Surgical treatment of traumatic fractures of the thoracic and lumbar spine: A systematic review

Introduction The treatment of traumatic thoracic and lumbar spine fractures remains controversial. To date no consensus exists on the correct choice of surgical approach and technique. Research question to provide a comprehensive up-to-date overview of the available different surgical methods and their quantified outcomes. Methods PubMed and EMBASE were searched between 2001 and 2020 using the term ‘spinal fractures’. Inclusion criteria were: adults, ≥10 cases, ≥12 months follow-up, thoracic or lumbar fractures, and surgery <3 weeks of trauma. Studies were categorized per surgical technique: Posterior open (PO), posterior percutaneous (PP), stand-alone vertebral body augmentation (SA), anterior scopic (AS), anterior open (AO), posterior percutaneous and anterior open (PPAO), posterior percutaneous and anterior scopic (PPAS), posterior open and anterior open (POAO) and posterior open and anterior scopic (POAS). The PO group was used as a reference group. Results After duplicate removal 6042 articles were identified. A total of 102 articles were Included, in which 137 separate surgical technique cohorts were described: PO (n = 75), PP, (n = 39), SA (n = 12), AO (n = 5), PPAO (n = 1), PPAS (n = 1), POAO (n = 2) and POAS (n = 2). Discussion and conclusion For type A3/A4 burst fractures, without severe neurological deficit, posterior percutaneous (PP) technique seems the safest and most feasible option in the past two decades. If needed, PP can be combined with anterior augmentation to prevent secondary kyphosis. Furthermore, posterior open (PO) technique is feasible in almost all types of fractures. Also, this technique can provide for an additional posterior decompression or fusion. Overall, no neurologic deterioration was reported following surgical intervention.


Introduction
Controversies remain on the optimal treatment of traumatic thoracic and lumbar spine fractures.Among different countries, regions, hospitals and even surgeons a wide range of treatment alternatives are seen, depending on their training and available materials (Vaccaro et al., 2016;Thomas et al., 2006;Abudou et al., 2013).In order to uniform the description, cluster fractures, and determine prognosis, a variety of classification systems have been introduced.Currently, the most commonly used system is the AO Spine Thoracolumbar Classification System, which incorporates radiological morphology and clinical factors (Vaccaro et al., 2013).However, as it does offer a uniform description for thoracic and lumbar fractures of the spine, it is not a treatment algorithm.
Depending on several clinical parameters, traumatic thoracic and lumbar spine fractures can be treated either surgical or conservative.Surgical treatment aims to reduce the fracture, prevent neurological deterioration, avoid further post-traumatic spinal deformity while maintaining spinal alignment and, if needed, perform a neural decompression.However, there is variety in literature regarding the optimal surgical approach and technique for different types of spine fractures.The decision depends on injury morphology, neurological status, integrity of the posterior ligamentous complex, available resources and, probably most importantly, surgeon expertise.Roughly, the types of surgical approach can be classified as anterior, posterior or combined.The choice depends on the required reduction, stabilization and neural decompression (Vaccaro et al., 2006).In 2004, Verlaan et al. published a thorough review that described the clinical relevance of the surgical treatment options for traumatic thoracic and lumbar fractures.Over the past two decades there have been various developments and less invasive techniques have become more mainstream.
Various meta-analysis and narrative reviews have been published on the treatment of traumatic thoracic and lumbar fractures of the spine (Xu et al., 2013;Lu et al., 2022;McAnany et al., 2016;Phan et al., 2015;Yi et al., 2006;Hughes et al., 2021;Sun et al., 2017;Tan et al., 2019;Kim et al., 2015;Pneumaticos et al., 2013).However, no comprehensive up-to-date overview is available of the different surgical methods, while using quantified outcome data.Therefore, the current authors have sought to provide an update on the 2004 review, using the same research team and more extensive methods (Verlaan et al., 2004).

Search strategy
A systematic literature search was performed in PubMed and EMBASE using the search term 'Spinal fractures' for articles published between 2001 and 2020.No language restrictions were imposed and a cross-reference check was performed (Appendix 1).For all missing fulltexts the library of our academic institution was consulted, and the authors were e-mailed with the request to provide the full text.

Study selection
Article screening was performed by four researchers (TV, MN, FP, SS).Inclusion criteria were: reporting adult subjects (>18 years of age); including more than 10 cases; a minimum of 12 months of follow-up; adequate description of the surgical procedure; using generally accepted outcome instruments; surgery performed within three weeks after trauma; including only thoracic or lumbar fractures.Studies that included non-traumatic fractures of the spine were excluded.The articles meeting the inclusion and exclusion criteria were extracted for fulltext analysis.

Data extraction
Data extraction was performed using a standardized form, simultaneously accessible for all authors (Google Spreadsheets, Google LLC, 2021/2022).Studies were categorized per surgical approach (posterior, anterior or combined) and sub-divided in one of the nine surgical technique groups.A description of the included studies is outlined in Appendix 1.

Data analysis
No quality assessment could be performed.Data was analyzed using the abovementioned surgical technique groups.Descriptive statistics were applied to summarize the distribution of values per group.The PO group was considered as the reference group, since it was found to be the most applied approach and included the largest number of studies and subjects.
To calculate the Cobb angle correction loss, the difference in Cobb angle between the directly postoperative Cobb angle and the Cobb angle at follow-up were calculated for each surgical technique group.Overall correction was defined as the difference between the pre-operative Cobb angle and the Cobb angle at follow-up.

Results
In total, 9151 articles were identified following the search strategy.A cross-reference check resulted in 107 additional articles.After duplicate removal, as well as title and abstract screening, 312 articles remained for further analysis.Of these articles, 209 were excluded due to various reasons (Fig. 1).Ultimately, 102 articles were analyzed of which 137 patient cohorts could be extracted that were treated with any of the surgical techniques.Five prospective controlled trials were found comparing the following.
• different time to surgery using traditional posterior pedicle screw constructs (Cengiz et al., 2008), • applying a traditional posterior pedicle screw construct using a less invasive posterior approach compared with an open approach (Chang et al., 2018), • the use of a traditional posterior pedicle screw construct compared with a group where posterior fusion was added (Jindal et al., 2012), • the use of a traditional posterior pedicle screw construct compared with a group with added laminar hooks (Korovessis et al., 2004), • posterior fusion with instrumentation, compared with anterior fusion with instrumentation (Wood et al., 2005).

Surgical characteristics
In the PO group, transpedicular spongioplasty using mainly autologous bone grafts, or a mixture with allograft or calcium sulfate was used in 19% of the subjects.In 58% of the subjects in the PO group, an additional decompression and fusion using various types of grafts was applied.Both the PO and PP groups showed small numbers of additional vertebroplasty or kyphoplasty (3% and 1%, respectively).Less fusion, grafts, decompression, and transpedicular spongioplasty were applied in the PP group.
The mean intra-operative blood loss was 430 ml in PO and 239 ml in PP.For SA, blood loss was only reported in two studies (0 ml and 450 ml) (He et al., 2013;Verlaan et al., 2005).The reported duration of surgery was lowest in the SA group, followed by the PP group and highest in the PO group (50, 95, and 134 min, respectively) (Table 3).

2007
) and 4% in AO group (Shin et al., 2020) had a complication of any sorts.

Follow-up
Reports of material failure were rare, yet occurred most following PO (2%), followed by PP (1%) and SA (0-1%).No material failure was found following AO and POAO.As shown in Table 4, reoperation rate was low in all groups, ranging from 0 to 2%.The Cobb angle preoperatively and at follow-up was widely reported (Table 5).The lowest overall correction was found following SA (1 • ), the other cohorts measured a minimum of 6 • .The change in ASIA score was reported in 53 cohorts (PO), 29 cohorts (PP), 10 cohorts (SA), 5 cohorts (AO), 2 cohorts (POAS) and 1 cohort (POAO).Overall, any surgical intervention led to an improvement of ASIA score or an unchanged ASIA score, with the exception of two cases of neurological deterioration.One patient in the PO group deteriorated from ASIA D to ASIA A and one in the PP group from ASIA D to ASIA B (Fig. 3) 725 .

Discussion
As this study shows, there has been a substantial number of publications on the surgical treatment of traumatic thoracic and lumbar spine fractures.However, the quality of the available evidence is low, with only a small number of randomized controlled trials (RCT).Various surgical treatment techniques have been compared in mostly monocenter retrospective cohort studies, with incomparable results due to a difference in follow-up time points and outcomes.This comprehensive systematic review quantifies the different outcomes data of all available studies concerning surgical treatment of traumatic thoracic and lumbar spine fractures over the past two decades.The quantity of the available data provides the best possible means to evaluate trends in surgical methods.
It is notable that the posterior percutaneous technique (PP) has been gaining in popularity in recent years.The posterior percutaneous technique is applied in similar rates and spinal regions in patients with type A3/A4 burst fractures and type B distraction injury fractures.It also measures comparable results to the posterior open technique (PO)  regarding kyphosis correction, with less blood loss and a shorter surgery duration.Overall, the postoperative immobilization (e.g., with a brace) period seemed slightly less compared to PO the hospital length of stay in the PP group was relatively longer.This could be explained due to the low amount of studies that reported on this subject and the differences between the patient population of the studies.The included studies by Phan et al. (2015) and Lu et al. (2022) compared PP with PO and reported a shorter hospital stay, shorter operation time, lower infection rates and better pain scores following surgery, while maintaining similar radiographic outcome data (Lu et al., 2022;Phan et al., 2015).However, PP requires a longer learning curve and may expose the surgeon to more radiation (Court and Vincent, 2012).Moreover, it is rarely used in combination with neural decompression (Wood et al., 2005;Fan et al., 2017;Kocis et al., 2020;Lee et al., 2013;Lyu et al., 2016;Pishnamaz et al., 2015;Vanek et al., 2014;Wang et al., 2017;Wang et al., 2013;Zhang et al., 2016;Zou et al., 2017;Afzal et al., 2008;Altay et al., 2007;Andress et al., 2002;Butt et al., 2007;Caruso et al., 2019;Caruso et al., 2018;Ding et al., 2021;He et al., 2013;Jiang et al., 2020;Kerschbaumer et al., 2019;Korovessis et al., 2008b;Korovessis et al., 2017;Korovessis et al., 2014;Li et al., 2016b;Manson et al., 2020; Padalkar and Mehta,   2017; Podet et al., 2020;Proietti et al., 2014;Sanli et al., 2019;Shao et al., 2020;Shin et al., 2020;Teyssedou et al., 2012;Trungu et al., 2019).Between 2009 and 2015 there was a rise in stand-alone vertebral augmentation (SA) use without fixation, mainly in Europe.Furthermore, only data of neurological intact patients that sustained a type A fracture was available.Advantages of the SA procedures include its minimal invasive approach, restoration of kyphosis, maintaining kyphosis correction and early pain reduction (He et al., 2013;Kerschbaumer et al., 2019;Teyssedou et al., 2012;Costa et al., 2009;Grelat et al., 2018;Hartmann et al., 2012;Hartmann et al., 2015;Klezl et al., 2011;Maestretti et al., 2014;Saget et al., 2014;Schmelzer-Schmied et al., 2009;Schulz et al., 2015;Verlaan et al., 2005).This procedure is mainly compared to conservative treatment with a brace, or a combination of posterior fixation and vertebroplasty or kyphoplasty.Three studies in the current review compared PP combined with SA to SA alone, and found that a combination of SA and PP had more favorable results (He et al., 2013;Kerschbaumer et al., 2019).This may be indicated when there is an A3 or A4 fracture with significant comminution or severely affected endplates.In this situation, protrusion of the intervertebral disc can occur through the endplates, possibly leading to failure of kyphosis correction (Teyssedou et al., 2012;Saget et al., 2014;Verlaan et al., 2005).The result of SA and PP combined was most favorable in a study by He et al. (2013), where only patients older than 65 years were included (He et al., 2013).This could be because of the increased benefit for elderly patients from anterior reinforcement and additional posterior reinforcement of the spine.The beneficial effect might be explained due to pre-existent sagittal-and coronal balance problems, possibly causing higher acting forces on the implants (Benoist, 2003).Furthermore, It should be noted that the risk of (osteoporotic) fractures may be higher in these patients, possibly resulting in different fracture characteristics limiting comparison to traumatic fractures (Waterloo et al., 2012).
Although studies with osteoporotic fractures were excluded, age-related loss of bone density may have affected these results.The Deutsche Gesellschaft für Orthopädie und Unfallchirurgie (DGOU) Osteoporotic Fracture Classification System makes a clear distinction in fracture assessment for osteoporotic fractures of the spine (Schnake et al., 2018).This distinction should be acknowledged in future studies, and different treatment strategies may be needed separately for non-osteoporotic vertebral fractures and osteoporotic vertebral fractures.
The studies on anterior approaches only addressed the open variant.A relatively large number of patients with neurological deficit and more severe fractures were included.One of the studies used a procedure called selective corpectomy, in which the vertebral body was spared as much as possible to reduce the risk of cage subsidence (Liang et al., 2017).Advantages mentioned in this study include a more thorough neural decompression and clearance of the spinal canal, the need of stabilization of fewer segments for the same strength of the construct compared to posterior fusion (Allain, 2011;Xu et al., 2013).However, it should be noted that no difference in neurological recovery was found.This is also found in the study by Xu et al., in which the open anterior (AO) approach was associated with longer operative time, similar neurological outcome, greater blood loss and higher costs than the posterior open (PO) approach for thoracolumbar fractures (Xu et al., 2013).Despite only few studies reported on complication rate, similar rate of complications were reported between both procedures (Shin et al., 2020;Liang et al., 2017;Oskouian et al., 2006;Smith et al., 2010;Yang et al., 2012).It is likely that the complication and immobilization rates is underreported in literature.

Implications
Due to the variability in study design of the included articles, these findings should be interpreted with care.The authors would like to  emphasize the importance of a uniform method to report the outcome of after treatment of traumatic thoracic and lumbar spine fractures.To standardize reporting, it is important to classify fractures according to the common and widely used classification systems, document neurological recovery according to a uniform system and report on relevant surgical characteristics (e.g., complication rate).Furthermore, the authors of this study would like to acknowledge the importance of including patient-reported outcomes for evaluating the success of surgical techniques during follow-up moment, besides using merely radiographic angles.Recently, the AO Spine PROST (Patient Reported Outcome Spine Trauma) was developed as the first patient-reported outcome measure specifically designed for spine trauma patients (Sadiqi et al., 2021;Sadiqi et al., 2017;Sadiqi et al., 2020).

Limitations
This systematic review has several limitations.There is a high probability of selection bias and publication bias, due to the retrospective nature and low sample sizes in most studies.However, due to the vast quantity of included studies, this study exemplifies a contemporary trend of the current practice of the different surgical techniques.Unfortunately, no meta-analysis could be performed due to the heterogeneity in methodology, follow-up, and outcome measures used in the included studies.Moreover, various important injury and patient characteristics-e.g., the frequency and severity of polytrauma injuries-were rarely reported.

Conclusion
In conclusion, this comprehensive systematic review describes and compares quantitative data regarding different surgical techniques for the treatment of traumatic thoracic and lumbar spine fractures.For type A3/A4 burst fractures without severe neurological deficit, posterior percutaneous (PP) technique seems the safest and most feasible option.If needed, PP can be combined with anterior augmentation in order to prevent secondary kyphosis in communitive fractures or severely affected endplates.Furthermore, posterior open (PO) technique is feasible in almost all types of fractures.Also, there seems to be particular use for this technique when additional posterior decompression or fusion is required, which cannot be achieved by more minimal invasive methods.Overall, no neurologic deterioration was reported following surgical intervention.Randomized controlled trials are imperative for a true comparison of surgical methods, and unfortunately were scarcely available.This systematic review exhibits a comprehensive overview of trends and clinical outcomes related to most common surgical treatment strategies in literature for patients with traumatic thoracic and lumbar spine fractures.

Declaration of competing interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: J. J. Verlaan reports a relationship with Philips Healthcare that includes: funding grants.J.J. Verlaan reports a relationship with SentryX that includes: board membership and equity or stocks.F.C. Oner reports a relationship with AO spine knowledge forum trauma that includes: board membership.F.C. Öner guest editor Brain and Spine If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 2 .
Fig. 2. Published cohorts per year and type of surgery.PO: Posterior open, PP: Posterior percutaneous, SA: Stand-alone vertebral body augmentation AO: Anterior open, PPAO: Posterior percutaneous and anterior open; PPAS: Posterior percutaneous and anterior scopic, POAO: Posterior open and anterior open, POAO Posterior open and anterior open, POAS: Posterior open and anterior scopic.

Table 1
General information on the study population.

Table 2
Pre-operative status of the study population.
PO: Posterior open, PP: Posterior percutaneous, SA: Stand-alone vertebral body augmentation, AO: Anterior open, PPAO: Posterior percutaneous and anterior open; PPAS: Posterior percutaneous and anterior scopic, POAO: Posterior open and anterior open, POAO Posterior open and anterior open, POAS: Posterior open and anterior scopic.a'All levels' includes thoracic, thoracolumbar and lumbar fractures, 'combination' includes a combination of two of the three regions.

Table 4
Complications at follow-up and immobilization duration.

Table 5
Cobb angle pre-operative, post-operative and at follow-up.Also reported are the correction lost and overall correction at final follow-up.