The efficacy and safety of anterior versus posterior approach for the treatment of thoracolumbar burst fractures: a systematic review and meta-analysis.

Background: There has always been controversy about the choice of anterior approach or posterior approach for the surgical approach of thoracolumbar burst fractures (TBFs). The aim of this study was to systematically evaluate the efficacy and safety of anterior and posterior approaches in the treatment of TBFs.
Methods: Multiple databases including PubMed, Excerpt Medica Database (Embase), Cochrane Library, and Web of Science (WOS) were used to search for relevant studies, namely full-text articles comparing the anterior versus posterior approach for the treatment of TBFs, which based on population, intervention, control, outcome, and study (PICOS) framework. Review Manager 5.4 was used to assess the effects of the results among selected studies. The risk of bias of the trials was assessed using the Newcastle Ottawa scale (NOS) and the Cochrane Collaboration's tool. Forest plots and funnel plots were also generated for the included articles.
Results: Finally, 723 patients were included in 13 studies which satisfied the eligibility criteria, funnel plots and Egger's test showed that there was no significant bias in the publications. There were no differences in terms of length of stay [mean difference (MD): -1.31, (-5.31, 2.69); P=0.52], hospitalization expenses [standardized mean difference (SMD): 1.26, (-0.38, 2.89); P=0.13], and return to work between the anterior approach and posterior approach. However, the posterior approach had the advantages of better Cobb angle correction [MD: 2.06, (0.17, 3.94); P=0.03], shorter operation time [MD: 58.29, (35.39, 81.18); P<0.00001], and lower estimated blood loss [MD: 185.92, (131.76, 240.07); P<0.00001]. Discussion: The posterior approach appeared to be superior to the anterior approach in the treatment of TBFs. However, more high-quality randomized controlled trials should be conducted to confirm the conclusions of this study and guide clinical decision-making. 2022 Annals of Translational Medicine. All rights reserved.

Introduction

The thoracolumbar vertebral body is a common site of spinal injury. It includes 4 vertebral bodies from the 11th vertebral body of the thoracic vertebrae to the 2nd vertebral body of the lumbar vertebrae (1,2). It is the connecting point between the lumbar spine and the relatively fixed thoracic spine, and is the part of the spine where the whole spine curve changes. The direction of the joint surface in this area changes (3). Due to these structural characteristics, vertebral bodies can easily fracture in this area (4). According to the literature, 90% of spinal fractures occur in the thoracolumbar region (5). Accidental falls are the most common cause of thoracolumbar vertebrae fractures, and other causes include traffic accidents and sports-related injuries. Thoracolumbar burst fractures (TBFs) caused by axial pressure account for about 10% to 20% of spinal fractures (6,7).

In TBF, the fracture usually occurs in the anterior and middle columns. Fractures usually lead to severe spinal instability, and even lead to acute or delayed neurological dysfunction (8,9). In these cases, surgery is usually required for treatment. The purpose of surgical intervention is nerve decompression, reconstruction of the vertebral body, and correction of angular deformity and stability (10). However, the method of surgery is still controversial (11).

There are three main methods currently used in surgery: anterior approach, posterior approach, and combination of anterior and posterior approaches (12-15). The anterior approach involves resection of the fractured vertebral body, direct decompression of the nerve, followed by anterior internal fixation (16). The posterior approach includes laminectomy and lateral approach combined with posterior indirect decompression (17). In anterior approach surgery, steel plates are used to fix the vertebral body, so as to obtain good decompression and firm fusion, but the risk of surgery is higher than that of posterior approach surgery (18). The posterior approach for TBF has the advantages of safety and a lower risk of damage to the lungs, internal organs, and vascular structures, and its technical requirements are relatively low (19).

Each surgical approach has its advantages and potential problems (20,21). For this reason, we collected related trials comparing anterior and posterior surgical techniques and performed a meta-analysis to explore the treatment of TBF with anterior and posterior approaches, so as to guide clinical practice for the surgical treatment of TBF. We present the following article in accordance with the PRISMA reporting checklist (available at https://atm.amegroups.com/article/view/10.21037/atm-22-903/rc).

Methods

Literature search strategy

Electronic databases including PubMed, Excerpt Medica Database (Embase), Cochrane Library, and Web of Science (WOS) were systematically searched for eligible studies from inception up to December 2021. We used the following search terms: (I) anterior approach; (II) posterior approach; (III) thoracolumbar burst fractures. All three search terms were combined with the Boolean operators “AND” and “OR” in the strategy. Full-text reviews were performed if the abstracts were insufficient for determining if the studies met the inclusion or exclusion criteria. Furthermore, the reference lists of the included studies and relevant review articles were screened for titles to identify additional appropriate articles for inclusion in this meta-analysis. Disagreements between the reviewers were resolved by discussion and consensus agreement.

Study selection

Studies were considered acceptable for inclusion in the review if they met the following criteria: (I) only patients diagnosed with TBF were included; (II) a comparison of surgical treatments from the anterior approach or posterior approach was performed; (III) indicators evaluating the efficacy between the anterior approach and posterior approach were used; (IV) only articles with full text available in English were selected. Studies were excluded if they were published in languages other than English, or were reviews, letters, or case reports without original data.

Data extraction

We developed a data extraction form by consensus. One researcher performed all of the data extraction, while the other researcher conducted independent verification. We extracted data of baseline characteristics, study design, author’s country, numbers of patients, patient characteristics (gender, age, fracture level), and follow-up time. The primary outcomes for evaluation were Cobb angle, operation time, length of hospital stay, estimated blood loss, records of return to work, hospitalization expenses, and complications.

Quality assessment

A subjective assessment of the methodological quality of the included studies was performed by two authors using the modified version of the Newcastle Ottawa scale (NOS) for non-randomized studies and the Cochrane Collaboration’s tool for randomized studies.

Statistical analysis

Meta-analysis was performed using Review Manager 5.4 provided by the Cochrane Collaboration. Continuous variables were expressed as mean difference (MD) or standardized mean difference (SMD) with 95% confidence interval (CI), and risk ratio (RR) was used for classification data. Heterogeneity across studies was assessed using Cochran Q and I2 statistics. The fixed effect model was applied in the absence of heterogeneity or for minor heterogeneity, and the random effect model was adopted for significant heterogeneity. Publication bias tests using funnel plots and Egger’s test were conducted in cases where there were ≥10 included studies.

Results

Search process

A total of 1,342 potentially relevant articles were identified by the searches. After removal of duplicates, 1,158 articles were identified. We found 85 titles after careful reading of the abstracts and full papers. In the further screening, 72 articles were excluded because of improper research and article type and insufficient data. Finally, 13 studies met our inclusion criteria and were selected for the present meta-analysis (22-34). The results of the search process were illustrated in a flowchart as shown in .

Figure 1

Schematic of the trial selection process.

Characteristics of the included studies

The detailed characteristics of these 13 eligible studies are summarized in . These studies contained 5 randomized controlled trials (RCTs) and 9 retrospective cohort studies (RCSs), which included 340 patients treated with the anterior approach and 383 patients treated with the posterior approach. The degree of fracture in patients included T11, T12, L1, and L2. All 13 articles were published from 1996 to 2020.

Table 1

Characteristics of eligible studies

Study	Study design	Country	No. of patients			Gender (M/F)			Age (years)			Level of fracture
			Anterior	Posterior		Anterior	Posterior		Anterior	Posterior		T11	T12	L1	L2
Danisa 1995	RCS	USA	16	27		11/5	19/8		35.4 [19–62]	36.8 [13–63]		0	9	28	5
Stancic 2001	RCT	Croatia	13	12		7/6	8/4		36 [18–53]	35 [16–60]		0	3	19	4
Wood 2005	RCS	USA	20	18		12/8	13/5		39.4±12.3	42.1±13.4		5	5	21	11
Hitchon 2006	RCS	USA	38	25		26/12	19/6		42±15	42±11		2	16	30	15
Sasso 2006	RCS	USA	40	13		29/11	10/3		–	–		–	–	–	–
Lin 2012	RCS	China	32	32		14/18	16/16		37.8±5.8	39.3±7.5		5	19	31	9
Chen 2012	RCT	China	18	18		10/8	12/6		38.7±5.9	40.2±7.35		5	11	15	5
Wu 2014	RCS	China	14	28		–	–		–	–		–	–	–	–
Wang 2015	RCT	China	22	23		14/8	15/8		37.2±11.4	40.5±13.5		0	9	30	6
Jiang 2019	RCT	China	40	40		–	–		–	–		–	–	–	–
Shin 2020	RCS	South Korea	22	24		17/5	13/11		46.9±12.4	43.4±12.4		0	4	25	17
Tan 2020	RCS	Australia	25	83		18/7	41/42		38.7±16.3	37.2±15.9		12	32	49	11
Yao 2020	RCT	China	40	40		–	–		–	–		–	–	–	–

RCS, retrospective cohort study; RCT, randomized controlled trial.

Results of quality assessment

The quality of the randomized studies and the risk of bias were assessed in accordance with the Cochrane Collaboration’s tools in the following six domains: selection, performance, detection, attrition, reporting, and other bias. The NOS was used to evaluate the methodological quality of non-randomized studies. The risk of bias of the non-RCTs is shown in . Except for one of the articles which was of moderate quality, the other 4 articles were of high quality. The risk of bias of the non-RCTs showed that all studies were rated over 7, which indicated no significant risk of bias ().

Table 2

Risk of bias of randomized controlled trials

Study	Random allocation	Allocation concealment	Blind method	Incomplete outcome data	Selective reporting of results	Other bias	Quality level
Stancic 2001	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk	High
Chen 2012	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk	High
Wang 2015	Unclear risk	Low risk	Low risk	Low risk	Low risk	Low risk	High
Jiang 2019	Low risk	Low risk	Low risk	Low risk	Unclear risk	Unclear risk	Moderate
Yao 2020	Low risk	Low risk	Low risk	Low risk	Low risk	Unclear risk	High

Table 3

Risk of bias of cohort studies

Study	Selection				Comparability of cohorts	Outcomes			Score*
	Representativeness of cohort	Selection of non-exposed cohort	Ascertainment of exposure	Outcome lacking at the beginning		Outcome assessment	Sufficient follow-up time	Follow-up adequacy
Danisa 1995	★	★	★	★	★★	☆	★	★	8
Wood 2005	★	★	★	★	★★	☆	★	★	8
Hitchon 2006	★	★	★	☆	★★	★	★	★	8
Sasso 2006	★	★	★	★	★☆	★	★	★	8
Lin 2012	★	★	★	★	★★	★	★	★	9
Wu 2014	★	★	★	★	★☆	☆	★	★	7
Shin 2020	★	★	★	★	★★	☆	★	★	8
Tan 2020	★	★	★	★	★★	☆	☆	★	7

*, the total score of NOS evaluation is 9 points; ★ represents that the item has obtained the score; ☆ represents that the item has not been scored. NOS, Newcastle Ottawa scale.

Results of the meta-analysis for outcomes

Cobb angle

Eleven trials evaluated the preoperative and postoperative Cobb angle change. High heterogeneity was found (P<0.00001; I2=98%). Consequently, a random effect model was applied. The posterior approach group showed a greater decrease in Cobb angle than the anterior approach group [MD: 2.06, (0.17, 3.94); Z=2.14, P=0.03] ().

Figure 2

Forest plot: anterior approach versus posterior approach for Cobb angle.

Operation time

Eleven studies involving 619 patients reported operation times. Pooled results from the random effect model showed that the anterior approach group experienced a longer operation time compared with the posterior approach group for TBF [MD: 58.29, (35.39, 81.18); Z=4.99, P<0.00001], and there was high heterogeneity among trials (P<0.00001, I2=95%) ().

Figure 3

Forest plot: anterior approach versus posterior approach for operation time.

Length of stay (LOS)

In terms of LOS, 6 studies involving 351 patients contributed to analysis. A random effect model was used due to the significant heterogeneity (P<0.00001, I2=94%). The pooled analysis showed that there was no difference in LOS between the anterior approach group and posterior approach group [MD: −1.31, (−5.31, 2.69); Z=0.64, P=0.52] ().

Figure 4

Forest plot: anterior approach versus posterior approach for length of stay.

Estimated blood loss

Eight studies involving 406 patients reported estimated blood loss. Pooled results from the random effect model showed that the anterior approach group had greater estimated blood loss compared with the posterior approach group for TBF [MD: 185.92, (131.76, 240.07); Z=6.73, P<0.00001], and there was high heterogeneity among trials (P<0.00001, I2=95%) ().

Figure 5

Forest plot: anterior approach versus posterior approach for estimated blood loss.

Hospitalization expenses

In terms of hospitalization expenses, 3 studies involving 125 patients contributed to analysis. A random effect model was used due to the significant heterogeneity (P<0.00001, I2=92%). The pooled analysis showed that there was no difference in hospitalization expenses between the anterior approach group and posterior approach group [SMD: 1.26, (−0.38, 2.89); Z=1.51, P=0.13] ().

Figure 6

Forest plot: anterior approach versus posterior approach for hospitalization expenses.

Return to work

Four studies on return to work showed mild heterogeneity in the consistency of the results (P=0.65; I2=0%). A fixed effect model was used for statistical analysis. No significant difference in return to work was found between the anterior approach group and posterior approach group [RR: 0.99, (0.78, 1.27); Z=0.05, P=0.96] ().

Figure 7

Forest plot: anterior approach versus posterior approach for return to work.

Complications

Nine studies on complications showed moderate heterogeneity in the consistency of the results (P=0.09; I2=42%). A fixed effect model was used for statistical analysis. The pooled analysis showed that the anterior approach group had a lower rate of complications than the posterior approach group [RR: 0.40, (0.27, 0.61); Z=4.40, P<0.0001] ().

Figure 8

Forest plot: anterior approach versus posterior approach for complications.

Publication bias

Publication bias was evaluated by visually inspecting funnel plots when at least 10 studies were included in the meta-analysis. Therefore, 2 funnel plots were produced for the outcomes of Cobb angle and operation time (). The plots showed some evidence of asymmetry, but Egger’s test for quantitative detection of publication bias showed that the bias was not statistically significant (Cobb angle, P=0.393; operation time, P=0.407).

Figure 9

Funnel plot of publication bias. (A) Cobb angle; (B) operation time.

Discussion

TBF are mostly caused by trauma. They are more common in young and middle-aged people, especially in men (5). They often damage the stability of the spine and can damage nerves. Most scholars believe that surgical treatment should be performed on TBF (35). The purpose is to relieve the compression of the spinal cord and restore the stability and nerve function of the spine to the greatest extent. However, the choice of approach is still clinically controversial (36,37).

The data in this study came from 8 non-RCTs and 5 RCTs, involving a total of 723 patients. Although the sample size included in this study was relatively small, we found that all included studies were of high quality and similar in baseline variables; therefore, we believed that the included studies were comparable. Through meta-analysis, we found that there was no difference in LOS [MD: −1.31, (−5.31, 2.69); P=0.52], hospitalization expenses [SMD: 1.26, (−0.38, 2.89); P=0.13], and return to work [RR: 0.99, (0.78, 1.27); P=0.96] between the anterior approach group and the posterior approach group, but the Cobb angle correction [MD: 2.06, (0.17, 3.94); P=0.03] of the posterior approach group was higher than that of the anterior approach group, the operation time [MD: 58.29, (35.39, 81.18); P<0.00001] was less than that of the anterior approach group, and the estimated blood loss [MD: 185.92, (131.76, 240.07); P<0.00001] of the posterior approach group was also lower than that of the anterior approach group. These findings highlight the advantages of the posterior approach over the anterior approach. However, we also found that the incidence of postoperative complications (RR: 0.40, (0.27, 0.61); P<0.0001) in the posterior approach group was higher than that in the anterior approach group.

In this study, the posterior approach could reduce bleeding, shorten the operation time, and effectively correct the Cobb angle, which is consistent with the findings of Tan et al. (21). However, their study found that the postoperative complications in the posterior approach group were higher than those in the anterior approach group, which may not be consistent with our results (38,39). The complicated anatomy and larger trauma of anterior approach surgery may lead to a higher rate of postoperative complications, though the results of our meta-analysis demonstrated contrasting findings. This may be due to the different definitions of postoperative complications in different studies (24,25). Some studies only included serious complications, while other studies included all complications during the operation. We will conduct a subgroup analysis of complications in future work.

Bolesta et al. demonstrated that the anterior approach was more appropriate for treating TBF with nerve damage and intact posterior ligaments (40). Anghel et al. found that there was no rigid standard for the anterior and posterior surgical approaches, but anterior approach surgery could effectively correct the angular deformity and maintain stability (41). Although anterior approach surgery can significantly relieve the compression on the front of the spinal cord and ensure satisfactory spinal fusion, it requires transthoracic and abdominal approaches due to the complex anatomy, and has high technical requirements (42). Therefore, the scope of development is relatively small. Posterior approach surgery is a traditional method for the treatment of TBF. It has simple anatomy, relatively superior technology, and results in short operation time, less trauma, and less bleeding. Early decompression surgery can significantly reduce secondary spinal nerve injuries and is widely used (43,44).

This meta-analysis had some limitations. Firstly, only 5 of the 13 included studies were RCTs, as most of the included studies were retrospective studies, which may reduce the reliability of the results. Secondly, the sample size was small, and the combined indicators of Cobb angle, operation time, LOS, and blood loss were highly heterogeneous, with varying degrees of selection bias, implementation bias, and measurement bias. Thirdly, the different follow-up times in each study might have affected our results.

Conclusions

In summary, for the correction of Cobb angle, the posterior approach may be a better choice, as it also had the characteristics of less bleeding and shorter operation time. However, the posterior approach may increase postoperative complications compared with the anterior approach. Due to the limitations of small sample size and lack of RCTs, the evidence for this meta-analysis is weak. Therefore, more high-quality RCTs are needed to confirm the conclusions of this study.

The article’s supplementary files as