Is Open Decompression Required in Thoracolumbar Spinal Trauma?

Article information

J Minim Invasive Spine Surg Tech. 2025;10(2):263-270

Publication date (electronic) : 2025 September 31

doi : https://doi.org/10.21182/jmisst.2025.02299

Veronica Parisi¹, Marianna Di Costanzo², Luigi Pintore², Chiara Caggiano¹, Gianpiero Iannuzzo¹, Giovanni Oliva¹, Claudio Schönauer^,1

¹Department of Neurosurgery, Cardarelli Hospital, Naples, Italy

²Division of Neurosurgery, Department of Neurosciences, Reproductive and Odontostomatological Sciences, University of Naples Federico II, Naples, Italy

Corresponding Author: Claudio Schönauer Department of Neurosurgery, Cardarelli Hospital, Azienda Ospedaliera di Rilievo Nazionale Antonio Cardarelli, Naples, 80131, Italy Email: claudio.schonauer@aocardarelli.it

Received 2025 April 29; Revised 2025 June 20; Accepted 2025 July 3.

Abstract

INTRODUCTION

Significant axial forces that brings about compression failure of the anterior and middle columns of the spine lead to burst fracture [1]. Most of them involve the thoracolumbar junction, which is uniquely due to its transitional anatomy and its anatomical location between the stiff kyphotic thoracic spine and the more mobile lordotic lumbar region. Typically, burst injuries are associated with some degree of spinal canal occlusion, which might result in neurologic deficits. The goals of surgery are to restore spinal stability through fracture stabilization and to improve functional outcomes by decompressing the neural elements. Many stable thoracolumbar fractures are non-surgically treated with external immobilization and early postoperative ambulation. However, the patient who exhibits spinal instability, progressive spinal deformity, or incomplete spinal cord injury is often an appropriate candidate for surgical intervention [2]. There is a great deal of controversy regarding the optimal surgical approach for treating a patient with a thoracolumbar burst fracture. The anterior or posterior surgery, as well as the short or long instrumentation are a matter of debate. Posterior instrumentation techniques are frequently used since they facilitate fracture reduction and subsequent arthrodesis. At the same time, indirect decompression of the spinal canal may be accomplished through distraction and ligamentotaxis, a process that effectively shifts the retropulsed bony fragments anteriorly, away from the neural structures. Ligamentotaxis in spinal fractures shows its efficacy by transferring the distraction force applied to pedicular screws to the posterior longitudinal ligament (PLL). As a result, it ensures the expansion of the spinal canal, the restoration of the vertebral height, and the correction of kyphosis [3].

This study aims to strengthen published data about ligamentotaxis procedures through posterior spinal instrumentation for thoracolumbar injuries. Moreover, the role of intraoperative imaging has been investigated for getting findings to perform surgery safely and efficiently. In particular, we sought to understand if intraoperative computed tomography (CT) scan might help to improve surgical approach and show whether open decompression have to be performed or not.

MATERIALS AND METHODS

A retrospective analysis of patients with traumatic spinal fracture was conducted at Cardarelli Hospital. The study period was between January 2023 and January 2025. Eligible for inclusion were those subjects who had undergone indirect decompression through ligamentotaxis procedures, using posterior spinal instrumentation for thoracolumbar single-segment compression injuries involving a single or both endplates along with a fractured posterior vertebral surface dislocated into the spinal canal (types A3 and A4 of AO Spine classification of thoracolumbar injuries). The PLL was undamaged in all patients. Hemodynamically unstable and paraplegic patients with complete cord transection, as well as those operated via anterolateral or combined approaches were not deemed suitable for the current study. Additionally, the ligamentotaxis procedure technique was considered contraindicated in patients with free bone fragments (“reverse cortical sign”). Surgery was performed no later than 24 hours after the accident. Seamless intraoperative imaging and registration with Loop-X, a Brainlab's (Germany) mobile imaging robot, was used to carry out surgical planning and navigation. Institutional electronic records system was searched for demographic, clinical, neurophysiological and neuroradiological data collection. A routine CT scan—performed at the time of hospital admission—was used to assess the fracture type and to define preoperative measurement of deformity which would be compared with the postoperative examination later. A magnetic resonance (MR) was also studied to evaluate the integrity of the spinal cord and posterior ligamentous complex. These CTs and MRs were undertaken preoperatively while patients underwent another CT scan while anaesthetized and in prone position to use during imaging-guided surgery (IGS). The last one CT scan was performed within 1 week postoperatively to prove surgical results. The primary outcome was the effectiveness of surgical procedure. Specifically, the canal widening, the restoration of corpus height, and the kyphosis correction were investigated using the PACS (picture archiving and communication system) software. The percentage of spinal canal compression was calculated using axial CT scans at the fractured level based on the formula: ([spinal canal area {SCA} above + below fracture]/2−SCA fracture)/([SCA above + below fracture]/2)×100; while the height of the fractured vertebral unit was evaluated from the bottom of the upper corpus to the bottom of the fractured corpus of the vertebra using sagittal CT scans [4]. Eventually, the kyphosis correction was measured between the superior endplate of the intact vertebra cephalad to the fracture and the inferior endplate of the vertebra caudal to the fracture (Cobb angle). Three raters, one board certified neurosurgeon (VP), a fifth and a fourth-year resident in Neurosurgery (MDC and PL, respectively), independently measured effectiveness outcomes. Surgery-related complications and details concerning quality of life, namely Oswestry Disability Index (ODI) and visual analogue scale (VAS) were deemed secondary outcomes. Length of surgical procedures was also reviewed. Written informed consent was obtained by the patient itself or, in case of subjects unable to provide it, by next of kin. The chance of a direct - open - decompression has always been emphasized at the informed consent in case indirect decompression was not reached after ligamentotaxis. The Guideline checklist of the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) statement for cohort studies was used [5].

Statistical analysis was performed using IBM SPSS Statistics ver. 30.0 (IBM Co., USA). Continuous data were expressed as mean±standard deviation with the interquartile range between brackets and compared with the Student t-test or with Mann-Whitney U-test when appropriate. Categorical variables were reported as number and percentage and compared with the Pearson chi-square with the Fisher exact test, if needed. Interobserver agreement was assessed using Cohen kappa (k). Results were considered statistically significant when p≤0.005.

All surgical procedures were performed under general anesthesia. Patients were positioned in prone position. The Loop-X Mobile Robotic Imaging BrainLab was used to instantly obtain intraoperative 3-dimensional (3D) images in order to navigate screws and instruments (Figure 1). After cleaning the skin and packing a sterile draping, 2-cm incision was performed on the midline 2 level above the target site to set registration reference. Subsequently, 1-cm skin incisions up to the fascia were made at vertebral pedicle according to IGS. Screws were inserted through the pedicles bilaterally. After contoured rod insertion, connecting caps were used to secure the rods within the heads of the pedicle screws. The caudal end of the construct was fixed, and distraction was applied throughout the fracture by a distractor to restore the vertebral body height, tension the annulus and PLL, and indirectly reduce any retropulsed fragments by ligamentotaxis. Then, the entire segment was locked in this lengthened position to maintain the physiological sagittal curve. Eventually, lumbar arthrodesis was performed placing synthetic bone graft through the incisions used for percutaneous surgery. Last CT scan was taken to show fixation device positioning and to decide if a decompressive laminectomy has to be performed. Wounds were copiously irrigated and closed in multiple layers.

Figure 1.

Operating theater and Loop-X Mobile Robotic Imaging (Brainlab, Germany).

RESULTS

Ten patients were included according to the eligibility criteria (Figure 2). The mean age at surgery was 36.1±17.9 years (29 [26–43.7] years). Sex was quite equally represented: 6 women and 4 men. The mechanisms of injury were motor vehicle accidents in 6 cases while 4 patients fell from a height. Four subjects had associated orthopedic fractures requiring surgery. Fracture localizations and types were determined according to AO spine classification of thoracolumbar injuries, with type A3 in 3 patients, type A4 in 7. The fractures were at T11 in 2 patients, at T12 in 2 patients, at L1 in 2 patients, at L2 in 2 patients, and at L4 in 2 patients. Measured preoperative spinal canal compression was 44%±15% (43 [36–46]); it became 22%±11% (22 [14–27]) postoperatively (p<0.001) (Figure 3). The vertebral unit height increased from 18.5±4.1 mm (18 [15.5–20.2] mm) to 31.9±3.5 mm (31 [29.5–34.0] mm) after surgery (p<0.001) (Figure 3). Greater widening was observed at the lumbar levels than at the dorsal ones, 16 mm versus 10 mm. Trauma-induced deformity was corrected in all patients: the initial kyphosis angle was 13.5°±2.4° (13.5° [11.5°–15.5°]) while the residual Cobb angle was 8°±2.5° (8° [6°–10°]) after surgery. A complete agreement among the 3 raters was obtained, Cohen kappa coefficient (κ) = 1. Effective reduction of the retropulsed fragment was obtained in all patients. There was no need of open surgery for decompression procedures according to intraoperative CT scan. The VAS scores decreased with preoperative and postoperative scores being 7.5±1.2 (7.5 [7–8.5]) and 2.9±1.5 (3 [2.4–4.0]), respectively (p=0.051). The mean ODI score was 28% (moderate disability)±13% (20% [20%–30%]) after discharge. No major complications such as iatrogenic dural or nerve root injury were observed. No infection occurred in the immediate postoperative period. There was no perioperative mortality or postoperative neurological deterioration. The mean postoperative hospital stay at Neurosurgery Department was 4.0±1.1 days (4 [3.5–4.5] days). Table 1 shows a comparison of outcomes among thoracolumbar spine fractures surgically treated with posterior instrumentation during the study period depending on whether IGS was used or not.

Figure 2.

Flowchart of enrolled patients. Eighty-nine patients underwent surgery for traumatic spinal fractures at our institution during the study period. Ten subjects met the inclusion criteria. Seventy-nine patients (70%) were excluded: 9 out of 89 (10%) due to anatomical localization of the injury in the cervical spine; and 70 out of 89 (79%) due to either an unfit surgical procedure and/or lack of intraoperative imaging.

Figure 3.

(A) Spinal canal area compression. Box-and-whisker plots showing the reduction in spinal canal area compression percentages among enrolled patients: 44%±15% (median [interquartile range], 43% [36%–46%]) versus 22%±11% (22% [14%–27%]) postoperatively. (B) Vertebral unit height. Box-and-whisker plots illustrating the increase in intervertebral unit height among enrolled patients: 18.5±4.1 mm (median [interquartile range], 18 [15.5–20.2] mm) to 31.9±3.5 mm (31 [29.5–34] mm). (C) Cobb's angle. Box-and-whisker plots showing reduction in kyphosis among enrolled patients: 13.5°±2.4° (median [interquartile range], 13.5° [11.5°–15.5°]) versus 8°±2.5° (8° [6°–10°]), p<0.001.

Table 1.

Comparison of clinical outcomes

DISCUSSION

Issues regarding the best treatment for thoracolumbar injuries with fractured posterior vertebral surface dislocated into the spinal canal (types A3 and A4 of AO Spine classification)—compression and burst fractures—has not yet been settled. Significant results have been reported both in the nonoperative and operative management [2,6,7]. Many studies have reported an improvement in canal dimension on subsequent CT scanning with conservative approach. As a matter of fact, the retropulsed fragments seems to gradually reabsorb up to 50% over a 2-year period [8-14]. Interestingly, the higher the amount of initial spinal canal compromise, the better the remodeling [9]. On the other hand, spontaneous canal remodeling, traditionally observed within 12 months after an injury, has been noted as early as 2- to 3-week postsurgery [15]. Eventually, no statistical difference has been found in the degree of spinal canal remodeling between patients treated conservatively and operatively [10-13]. Once surgery has been considered, the goal must be clearly defined to aid the selection of the appropriate procedure for achieving restoration of normal spinal anatomy together with neurological recovery. Indirect decompression techniques—such as posterior instrumented ligamentotaxis and anterior approaches followed by internal fixation—allow decompression of tissues, namely spinal cord and nerves, without resecting the compressing tissue. Thus, the risk of spinal nerve tissue injury decreases. Combined anterior and posterior approaches might also be considered. The current study focuses on the effect of computer-assisted ligamentotaxis in thoracolumbar burst fractures along with the advantages of intraoperative imaging for shortening the time of surgery. Our findings show results in line with previous studies. In particular, we observed a mean SCA decompression of 22% and 13.4 mm on average of vertebral unit height increase after surgery (Figures 4 and 5). Mueller et al. [16] described a postoperative SCA increase of approximately 10% in 36 patients with posterior distraction and ligamentotaxis while Crutcher et al. [17] demonstrated a 50% dropping in spinal canal stenosis with the same surgical approach in 13 cases. Although corpus height increases percentages and limits were not clearly stated in the literature, an increase of 20%–30% on average has been reported to date [18,19]. The ideal reduction maneuvers includes a combination of 5 mm of distraction with 6° of extension, or simply 10 mm of distraction, to decompress the spinal canal and intervertebral foramina in a cadaveric model [20]. Authors described the combination of 5 mm of distraction with 6° of extension using a straight rod with a sleeve while 10 mm of distraction using straight rods [18]. It is common knowledge that the rod needs to be cut a little longer than measured since distraction will lengthen the construct during surgery. The tendency towards greater vertebral unit height and canal widening after stabilization was apparent in our series although not significant clinically. However, a complete anatomic restitution of the spinal canal seem to be desirable, even if no valid correlations have been established between the restoration of collapsed height [21]. Of note, the advanced technology we used allowed us to visualize the spinal structure in 3 dimensions, so that surgical procedures were performed precisely and safely. As a matter of fact, our case series comes without any complications: no iatrogenic dural or nerve root injury was observed. The use of intraoperative CT imaging to view critical structures makes surgery precise, minimally invasive and safe providing reconstructed 3D images, which are more detailed than 2-dimensional fluoroscopic images, and add an axial view. Detailed 3D information may allow surgeons to identify misplaced pedicle screws more easily. However, it may also make surgeons repositioning suboptimal placed screws even when it is uncertain whether them would have caused any clinical symptoms postoperatively. Not just that, the intraoperative CT scan also helps surgeons to gather exact information about the bone fragment getting back in position. To date, the use of radiological methods, including intraoperative CT scan, has been only supposed to enhance surgical outcome. Yet notwithstanding these, it also seems like it may also help in making a decision which technique to used, in our experience. As a consequence, the effective reduction of a retropulsed fragment ensures to appreciate both when direct decompression through laminectomy is no longer needed or the need of laminectomy, that can be safely performed, when the SCA is quite enlarged after decompression [22]. The time of surgery is shorter for patients who underwent indirect decompression through ligamentotaxis using posterior spinal instrumentation and intraoperative cone-beam CT (Loop-X) control in our cohort; 82 minutes versus 120 minutes (p=0.011). Finally, we also found a mean postoperative hospital stay (at Neurosurgery Department for patient who suffered from other posttraumatic injuries) of 4 days for those thoracolumbar fractures treated with imagine-guided surgery (IGS) versus 7.1 days for those who underwent the same procedure without IGS since less tissue damage supports patients to recover faster after surgery. Notably, the same correlation is reflected on other parameter concerning the quality of life (postoperative VAS and ODI), see Table 1. Our study adds significant knowledge to existing literature by reporting great outcomes after indirect decompression through ligamentotaxis procedures using posterior spinal instrumentation for surgical planning, procedure navigation and intraoperative imaging control. However, its limitations include the design itself as well as the relatively small, enrolled population cohort that limits current statistical analysis. Moreover, the current lack of relevant the follow- up period cannot help with data indicating that the unreduced part of a retropulsed fragment could be resorbed gradually by remodeling. Surely, our protocol is permissible and provides insight for subsequent studies.

Figure 4.

(A) Preoperative sagittal computed tomography (CT) scan depicting the measurement of fractured vertebral unit height. (B) Preoperative axial CT scan showing the spinal canal area. (C) Preoperative axial CT scan displaying the Cobb angle at the burst fracture level. (D) Preoperative sagittal short tau inversion recovery (STIR) magnetic resonance imaging (MRI) demonstrating spinal cord compression. (E) Preoperative axial STIR MRI demonstrating spinal cord compression.

Figure 5.

(A) Postoperative computed tomography (CT) scans demonstrating spinal canal widening following posterior spinal instrumentation with distraction, ligamentotaxis, and bone cement augmentation. (B) Postoperative CT scans demonstrating apparent increase in vertebral unit height after posterior spinal instrumentation with distraction, ligamentotaxis, and bone cement augmentation. (C) Postoperative CT scans demonstrating Cobb angle correction at the burst fracture level after posterior spinal instrumentation with distraction, ligamentotaxis, and bone cement augmentation.

CONCLUSION

Indirect decompression techniques reduce retropulsed fragments improving both the degree of spinal canal compression and kyphotic angle. Furthermore, the use of intraoperative CT provides detailed information on bone fragment repositioning during surgery, aiding in the assessment of the need for decompressive laminectomy to further expand the spinal canal. If required, the procedure can be performed during the same surgical session, optimizing both operative time and overall costs.

Article information Continued