Bracing for Thoracolumbar Fractures
Bracing for Thoracolumbar Fractures
The treatment of burst fractures, or AO Type A3 fractures, is somewhat controversial within the literature. Under the Denis classification, this fracture type represents a 2-column injury and therefore can be considered unstable, but there exists a considerable body of literature that supports nonoperative treatment for these fractures. Burst fractures tend to occur at the thoracolumbar junction between T-11 and L-2. This area of the spine represents an interface between a relatively rigid segment of the thoracic spine, which is stabilized by the ribcage, compared with the relatively mobile segment of the lumbar spine. These fractures occur as often as 25,000 times a year in the US and represent approximately 10%–20% of all spinal fractures.
In 2003, Wood et al. published a prospective randomized study comparing operative to nonoperative treatment for thoracolumbar burst fractures. This study enrolled 53 patients, with 27 randomized to the nonoperative arm and 26 randomized to the operative arm. Inclusion criteria included an isolated burst fracture verified by plain radiographs as well as CT, no new neurological deficits, presentation less than 3 weeks from the time of injury, age between 18 and 66 years, and no other significant medical comorbidities or history of malignancy. Exclusion criteria included closed-head injuries, open vertebral fractures, neurological deficits, disruption of the posterior osteoligamentous complex, and osteoporotic insufficiency fractures. The visual analog scale (VAS), a modified Roland-Morris Disability Questionnaire (RMDQ), the Oswestry Disability Index (ODI), and the 36-Item Short Form Health Survey (SF-36) were used to assess clinical outcomes. Patients in the nonoperative arm were managed with either a body cast or a TLSO with the spine placed in hyperextension for 8 to 12 weeks. Patients in the operative arm were treated with either posterolateral short-segment fusion or anterior fusion with a fibular strut graft. Radiographic parameters included sagittal alignment and degree of canal compromise. Patients were followed up for as long as 2 years after enrollment into the study.
Overall, the randomized study by Wood et al. found no differences in radiographic findings between the operative and nonoperative treatment arms. The operative group had an average fracture kyphosis of 10.1° at the time of admission and 13° at the final follow-up. The average canal compromise was 39% on admission, and it improved to 22% at the final follow-up examination. In the nonoperative group, the average kyphosis was 11.3° at the time of admission and 13.8° at the final follow-up. Average canal compromise at the time of admission was 34% and improved to 19% at the final follow-up. For clinical outcomes the average VAS scores were similar between both groups. On the RMDQ, the nonoperative group reported less disability at the final follow-up examination, with a score of 8.16 compared with 3.9 for the operative group (p = 0.02). The average ODI also favored the non-operative group, with a final mean score of 20.75 for the operative group and 10.7 for the nonoperative group. The SF-36 scores demonstrated significant differences with respect to physical function (p = 0.002) and role (p = 0.003), favoring nonoperative treatment. No differences in return to work were noted between the 2 groups.
The authors also assessed hospital costs, noting that on average the operative group had longer hospital stays (7.9 days for the nonoperative group compared with 10.7 days for the operative group). A total hospital charge for the operative group was approximately $49,063 versus $11,264 for the nonoperative group. The authors concluded that surgical treatment of thoracolumbar burst fractures showed no substantial benefit over nonoperative treatments. This study represented the first prospective randomized study comparing bracing to surgical treatment for burst fractures.
Post et al. published a series of papers regarding outcomes of nonoperatively treated patients with follow-up of as long as 10 years. Their initial study with 5 years of follow-up consisted of 38 patients, whereas their more recent study included 50 patients who had both 4-year and 10-year data available. The initial study used the VAS, SF-36, and RMDQ as validated outcome measures. This study also included a dynamic lifting test as well as an ergometry test to assess functional capacity as well. They found that 37% of the patients were not able to perform the dynamic lifting test within the normal range. In addition, 40.9% of the patients scored below the lowest normal values on the ergometry test. The mean RMDQ score at 5 years was 5.2, while the mean VAS score was 79. Overall, patients were mildly disabled compared with norms but not significantly so. Finally, only 10% of patients had stopped working as a result of back problems related to their fracture. In their follow-up study, the VAS and RMDQ results were the only outcomes available for analysis. The authors reported at midterm follow-up (average 4.3 years) VAS and RMDQ scores of 74.5 and 4.9, respectively. At the long-term follow-up (average 9.8 years), the VAS and RMDQ scores were 72.5 and 4.7, respectively. The authors concluded that patients treated nonoperatively reached a steady state of disability at approximately the 4-year mark without any further deterioration of function or worsening disability. In addition, the authors believed that given their results, nonoperative treatment was a viable option in treating type A fractures. While not as robust as the randomized trial of Wood et al., the studies published by Post and colleagues did show durability of functional outcome with bracing of thoracolumbar fractures up to a 10-year time point.
The need for bracing or no bracing was also evaluated by Bailey et al. in a prospective randomized multi-center trial. This study included patients with AO Type A3 burst fractures with a kyphotic deformity less than 35°, who were neurologically intact, 16 to 60 years of age, and within 3 days of injury. A total of 96 patients were enrolled in the study, with 47 randomized to treatment with TLSO and 46 randomized to treatment with no orthosis. Patients in the TLSO group were to remain on strict bed rest prior to being fitted with a TLSO and were instructed to wear the brace for at least 10 weeks. Clinical outcomes measured included the RMDQ, SF-36, and VAS. Radiographic outcomes included the Cobb angle of sagittal alignment. Follow-up intervals included 3 months, 12 months, and 24 months after injury.
Five patients required surgical treatment: 2 patients had severe radicular symptoms upon mobilization, and 3 other patients had severe mechanical back pain necessitating stabilization. With these patients excluded, 48 patients received no orthosis and 43 received TLSO. The authors found that at all time points there were no statistical differences detected in RMDQ, SF-36, and VAS scores between treatment groups. The mean RMDQ score was 9.8 for the no orthosis group and 8.7 for the TLSO group at 2-year follow-up. The mean physical component scores (PCSs) of the SF-36 were 36.6 and 39.1 at 2 years for the no orthosis and TLSO groups, respectively, while the mean mental component scores (MCSs) were 50.8 and 52.2 for the no orthosis and TLSO groups, respectively. Average kyphosis was also similar between groups at admission (no orthosis, 14°, and TLSO, 15°), discharge (20° and 18°, respectively), 6 weeks (21° and 21°, respectively), and all other time points. Based on these results, the authors concluded that thoracolumbar burst fractures are inherently stable fractures that can be managed without orthosis.
However, not all clinical studies support the use of bracing over surgery. Siebenga et al. published a multi-center randomized trial in 2006 comparing bracing with posterior short-segment fusion for Type A thoracolum-bar fractures. The study enrolled 34 patients, with 16 randomized to nonsurgical therapy and 18 randomized to surgical therapy. The nonsurgical patients were maintained on bed rest for a minimum of 5 days, and then they received Jewett hyperextension orthosis and were instructed to wear the brace for 3 months. Patients in the surgical arm received short-segment posterior stabilization. Patients were again followed up for up to 2 years. The RMDQ and VAS scores were used to assess clinical outcomes. Sagittal Cobb angles were used for radiographic outcomes.
At the time of final follow-up, kyphosis was 19.8° in the nonoperative group and 8.4° in the surgical group (p < 0.0001). For disability, the mean RMDQ score in the nonsurgical group was 8.9, and for the surgical group it was 3.1 (p = 0.030). For pain assessment, the mean VAS score in the nonsurgical group was 61 compared with 81 for the surgical group (p = 0.020). Five patients randomized to surgery experienced complications, 2 of whom had to return to the operating room for wound infection treatment and hardware revision. Three patients in the nonsurgical group had complications, and 1 of the 3 patients developed a severe psychological disorder as a result of her injury. Only 38% of patients in this study who were managed nonoperatively were able to return to work, while 85% of those who were managed surgically were able to return to work. The authors concluded that short-segment posterior fusion was superior to bracing for treatment of burst fractures.
When interpreting the study of Siebanga et al., it should be noted that the RMDQ scores for patients who were treated with bracing were similar to those in the Post et al. and Bailey et al. studies. However, the braced cohort in the study by Wood et al. suffered significantly less disability than the patients treated with bracing in the study by Siebenga et al. In addition, return to work as well as VAS scores in the Siebenga et al. study were also poorer than those reported for the other studies presented above. Radiographically, Siebenga et al. had superior results to those presented by Wood et al. when both surgical cohorts in the 2 studies were compared. Whether the difference in radiographic results has any bearing on the differing clinical results between the 2 studies is important to consider, given the importance of sagittal balance as a driver for pain and disability in patients with adult deformity. Nevertheless, there is a large amount of evidence that supports the use of bracing for the management of burst-type thoracolumbar fractures, and in some studies bracing appears superior to operative treatment. Therefore, bracing should be considered in most neurologically intact patients with burst fractures.
Controversies in Treatment
The treatment of burst fractures, or AO Type A3 fractures, is somewhat controversial within the literature. Under the Denis classification, this fracture type represents a 2-column injury and therefore can be considered unstable, but there exists a considerable body of literature that supports nonoperative treatment for these fractures. Burst fractures tend to occur at the thoracolumbar junction between T-11 and L-2. This area of the spine represents an interface between a relatively rigid segment of the thoracic spine, which is stabilized by the ribcage, compared with the relatively mobile segment of the lumbar spine. These fractures occur as often as 25,000 times a year in the US and represent approximately 10%–20% of all spinal fractures.
In 2003, Wood et al. published a prospective randomized study comparing operative to nonoperative treatment for thoracolumbar burst fractures. This study enrolled 53 patients, with 27 randomized to the nonoperative arm and 26 randomized to the operative arm. Inclusion criteria included an isolated burst fracture verified by plain radiographs as well as CT, no new neurological deficits, presentation less than 3 weeks from the time of injury, age between 18 and 66 years, and no other significant medical comorbidities or history of malignancy. Exclusion criteria included closed-head injuries, open vertebral fractures, neurological deficits, disruption of the posterior osteoligamentous complex, and osteoporotic insufficiency fractures. The visual analog scale (VAS), a modified Roland-Morris Disability Questionnaire (RMDQ), the Oswestry Disability Index (ODI), and the 36-Item Short Form Health Survey (SF-36) were used to assess clinical outcomes. Patients in the nonoperative arm were managed with either a body cast or a TLSO with the spine placed in hyperextension for 8 to 12 weeks. Patients in the operative arm were treated with either posterolateral short-segment fusion or anterior fusion with a fibular strut graft. Radiographic parameters included sagittal alignment and degree of canal compromise. Patients were followed up for as long as 2 years after enrollment into the study.
Overall, the randomized study by Wood et al. found no differences in radiographic findings between the operative and nonoperative treatment arms. The operative group had an average fracture kyphosis of 10.1° at the time of admission and 13° at the final follow-up. The average canal compromise was 39% on admission, and it improved to 22% at the final follow-up examination. In the nonoperative group, the average kyphosis was 11.3° at the time of admission and 13.8° at the final follow-up. Average canal compromise at the time of admission was 34% and improved to 19% at the final follow-up. For clinical outcomes the average VAS scores were similar between both groups. On the RMDQ, the nonoperative group reported less disability at the final follow-up examination, with a score of 8.16 compared with 3.9 for the operative group (p = 0.02). The average ODI also favored the non-operative group, with a final mean score of 20.75 for the operative group and 10.7 for the nonoperative group. The SF-36 scores demonstrated significant differences with respect to physical function (p = 0.002) and role (p = 0.003), favoring nonoperative treatment. No differences in return to work were noted between the 2 groups.
The authors also assessed hospital costs, noting that on average the operative group had longer hospital stays (7.9 days for the nonoperative group compared with 10.7 days for the operative group). A total hospital charge for the operative group was approximately $49,063 versus $11,264 for the nonoperative group. The authors concluded that surgical treatment of thoracolumbar burst fractures showed no substantial benefit over nonoperative treatments. This study represented the first prospective randomized study comparing bracing to surgical treatment for burst fractures.
Post et al. published a series of papers regarding outcomes of nonoperatively treated patients with follow-up of as long as 10 years. Their initial study with 5 years of follow-up consisted of 38 patients, whereas their more recent study included 50 patients who had both 4-year and 10-year data available. The initial study used the VAS, SF-36, and RMDQ as validated outcome measures. This study also included a dynamic lifting test as well as an ergometry test to assess functional capacity as well. They found that 37% of the patients were not able to perform the dynamic lifting test within the normal range. In addition, 40.9% of the patients scored below the lowest normal values on the ergometry test. The mean RMDQ score at 5 years was 5.2, while the mean VAS score was 79. Overall, patients were mildly disabled compared with norms but not significantly so. Finally, only 10% of patients had stopped working as a result of back problems related to their fracture. In their follow-up study, the VAS and RMDQ results were the only outcomes available for analysis. The authors reported at midterm follow-up (average 4.3 years) VAS and RMDQ scores of 74.5 and 4.9, respectively. At the long-term follow-up (average 9.8 years), the VAS and RMDQ scores were 72.5 and 4.7, respectively. The authors concluded that patients treated nonoperatively reached a steady state of disability at approximately the 4-year mark without any further deterioration of function or worsening disability. In addition, the authors believed that given their results, nonoperative treatment was a viable option in treating type A fractures. While not as robust as the randomized trial of Wood et al., the studies published by Post and colleagues did show durability of functional outcome with bracing of thoracolumbar fractures up to a 10-year time point.
The need for bracing or no bracing was also evaluated by Bailey et al. in a prospective randomized multi-center trial. This study included patients with AO Type A3 burst fractures with a kyphotic deformity less than 35°, who were neurologically intact, 16 to 60 years of age, and within 3 days of injury. A total of 96 patients were enrolled in the study, with 47 randomized to treatment with TLSO and 46 randomized to treatment with no orthosis. Patients in the TLSO group were to remain on strict bed rest prior to being fitted with a TLSO and were instructed to wear the brace for at least 10 weeks. Clinical outcomes measured included the RMDQ, SF-36, and VAS. Radiographic outcomes included the Cobb angle of sagittal alignment. Follow-up intervals included 3 months, 12 months, and 24 months after injury.
Five patients required surgical treatment: 2 patients had severe radicular symptoms upon mobilization, and 3 other patients had severe mechanical back pain necessitating stabilization. With these patients excluded, 48 patients received no orthosis and 43 received TLSO. The authors found that at all time points there were no statistical differences detected in RMDQ, SF-36, and VAS scores between treatment groups. The mean RMDQ score was 9.8 for the no orthosis group and 8.7 for the TLSO group at 2-year follow-up. The mean physical component scores (PCSs) of the SF-36 were 36.6 and 39.1 at 2 years for the no orthosis and TLSO groups, respectively, while the mean mental component scores (MCSs) were 50.8 and 52.2 for the no orthosis and TLSO groups, respectively. Average kyphosis was also similar between groups at admission (no orthosis, 14°, and TLSO, 15°), discharge (20° and 18°, respectively), 6 weeks (21° and 21°, respectively), and all other time points. Based on these results, the authors concluded that thoracolumbar burst fractures are inherently stable fractures that can be managed without orthosis.
However, not all clinical studies support the use of bracing over surgery. Siebenga et al. published a multi-center randomized trial in 2006 comparing bracing with posterior short-segment fusion for Type A thoracolum-bar fractures. The study enrolled 34 patients, with 16 randomized to nonsurgical therapy and 18 randomized to surgical therapy. The nonsurgical patients were maintained on bed rest for a minimum of 5 days, and then they received Jewett hyperextension orthosis and were instructed to wear the brace for 3 months. Patients in the surgical arm received short-segment posterior stabilization. Patients were again followed up for up to 2 years. The RMDQ and VAS scores were used to assess clinical outcomes. Sagittal Cobb angles were used for radiographic outcomes.
At the time of final follow-up, kyphosis was 19.8° in the nonoperative group and 8.4° in the surgical group (p < 0.0001). For disability, the mean RMDQ score in the nonsurgical group was 8.9, and for the surgical group it was 3.1 (p = 0.030). For pain assessment, the mean VAS score in the nonsurgical group was 61 compared with 81 for the surgical group (p = 0.020). Five patients randomized to surgery experienced complications, 2 of whom had to return to the operating room for wound infection treatment and hardware revision. Three patients in the nonsurgical group had complications, and 1 of the 3 patients developed a severe psychological disorder as a result of her injury. Only 38% of patients in this study who were managed nonoperatively were able to return to work, while 85% of those who were managed surgically were able to return to work. The authors concluded that short-segment posterior fusion was superior to bracing for treatment of burst fractures.
When interpreting the study of Siebanga et al., it should be noted that the RMDQ scores for patients who were treated with bracing were similar to those in the Post et al. and Bailey et al. studies. However, the braced cohort in the study by Wood et al. suffered significantly less disability than the patients treated with bracing in the study by Siebenga et al. In addition, return to work as well as VAS scores in the Siebenga et al. study were also poorer than those reported for the other studies presented above. Radiographically, Siebenga et al. had superior results to those presented by Wood et al. when both surgical cohorts in the 2 studies were compared. Whether the difference in radiographic results has any bearing on the differing clinical results between the 2 studies is important to consider, given the importance of sagittal balance as a driver for pain and disability in patients with adult deformity. Nevertheless, there is a large amount of evidence that supports the use of bracing for the management of burst-type thoracolumbar fractures, and in some studies bracing appears superior to operative treatment. Therefore, bracing should be considered in most neurologically intact patients with burst fractures.
Source...