Risk of vertebral artery injury associated with subaxial cervical pedicle screw malposition: cadaveric evidence and clinical relevance
Article information
Abstract
Study Design
Experimental cadaveric anatomical study.
Purpose
This study aims to test the safe zone hypothesis for preventing vertebral artery injury as a notable complication of subaxial cervical spine pedicle screw malpositioning and determine the critical threshold at which screw malposition causes vertebral artery injury.
Overview of Literature
The pedicle screw fixation is the strongest biomechanical stabilization approach in the posterior cervical region. Although it is associated with potential complications like vertebral artery injury, these complications are rarer than anticipated. Accordingly, several authors have proposed a “safe zone” within which even a malpositioned screw may displace and not injure the artery.
Methods
Five human cadavers were subjected to transpedicular screw fixation for creating lateral pedicle violation. Pre- and post-procedural anatomical measurements were obtained using computed tomography scans. Postoperatively, vertebral arteries were dissected to assess damage macroscopically and histopathologically.
Results
A significant correlation was observed between the safety margin (the combined distance of the vertebral artery from the lateral and medial walls of the transverse foramen) and the intraforaminal area occupied by screws for causing vertebral artery damage (p<0.001). The critical threshold for causing vertebral artery damage is exceeding this safety margin.
Conclusions
This study experimentally validated the “safe zone” hypothesis, confirming its presence and effectiveness. In cases of screw breach, the decision for revision surgery should be based on the patient’s clinical condition and breach status (vessel displaced or injured), rather than the screw position alone. However, if a vertebral artery injury is suspected, acute intervention should be initiated promptly.
Introduction
Pedicle screw fixation in the subaxial cervical spine was first described by Abumi et al. [1] in 1994. Initially used for traumatic cervical spine fractures, the indications for pedicle screw fixation have since been expanded to include nontraumatic lesions, craniocervical junction reconstruction, correction of cervical kyphosis, and spondylotic myelopathy with cervical deformity or instability [2–5], primarily because of its biomechanical ability to offer the strongest stabilization in the posterior cervical region with the lowest risk of screw loosening [6–8].
Despite its numerous biomechanical advantages, the pedicle screw is used less commonly compared to the lateral mass screw owing to the risk and severity of potential complications with the former, particularly owing to the variations in the size and angle of the cervical pedicle [9]. The anatomical differences, along with the smaller size of the cervical pedicle compared to thoracic and lumbar pedicles, amplify the risk of screw malpositioning in the cervical spine [10]. Previous studies have shown that approximately 80% of screw malpositioning cases occur laterally because the lateral pedicle wall is thinner and the paraspinal muscles limit excursion in the medial angle [11–13]. This lateral breach is crucial in the cervical spine as it can lead to vertebral artery injury, potentially causing pseudoaneurysms, fistulae, emboli, and infarcts in the cerebral, cerebellar, or brainstem regions supplied by the posterior circulation, with a variety of clinical symptoms and even mortality, depending on the location of the infarct [14].
Interestingly, despite the high breach rates, the incidence of vertebral artery injury owing to cervical pedicle screw malpositioning is much lower than expected. Soliman et al. [15] reported a 0.2% incidence of vertebral artery injury with pedicle screw fixation, making it as safe as the lateral mass technique in terms of vascular injury. The lower-than-expected rates of vertebral artery complications from screw breaches into the transverse foramen have led to the concept of a “safe zone” in the cervical spine. Several authors have defined this zone in different anatomical spaces and angles [12,16–20], suggesting that the vertebral artery can withstand a certain degree of screw malposition in the absence of injury; however, there is a lack of a universal definition and experimental studies supporting this hypothesis of a safe zone.
The present study aims to establish an objective criterion for defining the safe zone and the degree of arterial malposition that the vertebral artery can withstand without injury during cervical spine surgery.
Materials and Methods
Study design and subjects
This experimental cadaveric anatomical study was conducted on five male human cadavers with intact head and neck regions. The cadavers were fixed with formaldehyde, and the vascular structures were filled with silicone. Macroscopic examination prior to the study confirmed that the cadavers had not undergone any cervical spine interventions.
Ethical considerations
This study was conducted in compliance with the principles of the Declaration of Helsinki. The study’s protocol was reviewed and approved by the Ethics Committee of Marmara University Faculty of Medicine Clinical Research (protocol number: 09.2023.242). The need for informed consent was waived due to the study design.
Pre-procedural imaging
Before intervention, the cadavers were subjected to 1-mm thickness computed tomography (CT) scans and magnetic resonance imaging (MRI) (T1 and T2 sequences) to assess cervical anatomy, vertebral artery status, transverse foramen structure, and any previous surgical interventions. Fifteen cervical spine levels (30 pedicles) from the five cadavers were included based on predefined criteria.
Subsequently, three-dimensional (3D) reconstructions of the preprocedural CT images were performed in the axial, sagittal, and coronal planes, and the vertebral artery area (VAA) and transverse foramen area (TFA) were calculated using closed polygon measurement on axial slices parallel to the pedicle. Anatomical measurements, including the coronal and sagittal lengths of the vertebral artery and transverse foramen, the shortest distance of the vertebral artery from the pedicle, the artery-to-medial border of transverse foramen distance (AFMD), and artery-to-lateral border of the transverse foramen distance (AFLD), were taken at the midpoints. Additionally, pedicle thickness at the level of the isthmus was measured on the 3D-CT images. The vertebral artery occupation ratio (AOR) was calculated by comparing the vertebral artery area and the transverse foramen area to determine the area occupied by the vertebral artery in the transverse foramen (Fig. 1).
Schematic representation of anatomical measurements and angles. (A) Coronal length of artery (CLA), sagittal length of artery (SLA), artery-pedicle closest distance (APD), artery medial to foramen (AFMD), and artery lateral to foramen (AFLD). (B) Pedicle width (PW), sagittal length of transvers foramen (SLF), and coronal length of transvers foramen (CLF). (C) a: pedicle angle (PA), b: angle through the medial foramen (FMA), and c: angle through the medial artery (AMA).
As suggested by Sharma et al. [17], the safety margin (SM) was defined as the sum of AFMD and AFLD, representing the length of the free space within the transverse foramen in the coronal plane. However, unlike Sharma et al. [17], who used the formula “πr2” for VA and TFA based on the assumption that the vertebral artery and transverse foramen are circular, our study used more specific measurements to accommodate the noncircular morphology of these structures.
Additionally, the transverse pedicle angle (PA), the angle of passing through the medial border of the transverse foramen (FMA), and the angle of the line passing through the medial border of the vertebral artery (AMA) were measured on axial sections parallel to the pedicle (Fig. 1).
Screw placement
All cadavers, except one that was dissected to illustrate the surrounding anatomy, were opened with a midline incision appropriate for surgery, exposing the spinal elements and the V3 section of the vertebral arteries. As the silicone material in the vertebral artery could not provide sufficient pressure and resistance during screw placement, an experimental setup was created to simulate in vivo conditions.
The vertebral artery was cut at the appropriate distal section after exiting the C2 transverse foramen, and the silicone material was flushed out with saline. An Angiocath (Becton, Dickinson and Company, Franklin Lakes, NJ, USA) was attached to the distal and proximal parts of the V2 portion of the vertebral artery, converting it into a closed system. A saline infusion set, placed at a height of 180 cm, was connected to the proximal Angiocath to maintain a constant pressure of 180 cmH2O (132 mm Hg) in the V2 portion. Before screw insertion, the vertebral artery was observed for 5 minutes with this system to ensure its integrity.
After exposing the lateral mass and identifying the pedicle entry point, the cortical surface was opened using a high-speed burr. Polyaxial titanium cervical pedicle screws (3.5 mm in diameter, 14 mm in length) were inserted using the freehand technique as described by Abumi et al. [1]. The entry point was located slightly lateral to the midpoint of the lateral mass and just inferior to the superior articular facet; the screws were advanced in a mediolateral direction with a trajectory parallel to the superior endplate in the sagittal plane. Unlike the original technique by Abumi et al. [1], we used a deliberately reduced medial angulation relative to the native pedicle axis to reproduce controlled lateral breaches for experimental purposes. A blunt pedicle probe was used to prepare pilot channels; tapping was performed prior to screw insertion to minimize pedicle fractures. All screws had a constant thread pitch and cylindrical core design to standardize biomechanical interaction within the pedicle. All insertions were performed by a single spine surgeon to avoid operator variability.
Post-procedural imaging and measurements
After screw placement, the cadavers were again subjected to CT imaging with 1-mm axial slices; the CT images were then processed using three-plane and 3D reconstruction. If the screw had entered the transverse foramen, the angle of screw insertion to the midline and the extent of screw violation (ASV) were calculated from the 3D images corresponding to the point where the screw presented a deviation farthest laterally from the pedicle (Fig. 2).
The relationship between the vertebral artery and the pedicle screw. (A) Measurement of the amount of screw violating the foramen on a three-dimensional (3D) image. Part violating the foramen (purple) and amount of transvers foramen violation (yellow line). (B) Transverse foramen with the lateral wall opened and the relationship of the neurovascular structures to the screw. (C) 3D image of the same region. Vertebral artery (red) and cervical nerves (yellow).
All CT-based morphometric measurements were taken by a single surgeon with experience in cervical pedicle morphometry. To assess their consistency, all measurements were repeated after a 2-week interval, with the surgeon blinded to their previous measurements.
Vertebral artery resection and histopathological analysis
After imaging, soft-tissue dissection was performed to expose the lateral aspect of the cervical spine. The lateral wall of the transverse foramen was excised using a high-speed drill and Kerrison’s rongeur to access the vertebral arteries in the foramen (Fig. 2); vertebral arteries at each level were excised separately, and tissue samples were collected. Arteries with macroscopic full-thickness injury at the site of screw contact were recorded as “injured,” whereas those without such injury were placed in 10% formaldehyde for histopathologic examination.
The vascular specimens reserved for histopathologic examination were sectioned into 2-μm slices; stained with hematoxylin-eosin, Masson’s trichrome, and Verhoeff’s Van Gieson stain; and examined for damage to the vessel wall structures and layers.
Statistical analysis
All analyses were performed using IBM SPSS ver. 29.0. (IBM Corp., Armonk, NY, USA). Descriptive statistics for continuous variables were reported as mean±standard deviation, whereas categorical variables were presented as frequency and percentages. Normality of the data distribution was assessed using the Shapiro-Wilk test.
Pearson’s or Spearman’s correlation coefficients were used for assessing normally or non-normally distributed variables, respectively. For categorical variables, Pearson’s chi-square test was applied to 2×2 tables with expected observations of ≥5, Fisher’s exact probability test to tables with ≤5 observations, and Fisher–Freeman–Halton test to R×C tables with <5 expected observations. Additionally, Cramer’s V was calculated for assessing the correlation between categorical data.
An independent-samples t-test was employed for comparing two groups with normally distributed data. One-way analysis of variance (ANOVA) was used to compare normally distributed multiple independent groups, followed by Tukey’s test for post hoc analysis. Intraobserver reliability was calculated using the intraclass correlation coefficient (ICC). For all analyses, the significance level (α) was set at 0.05.
Results
A total of 30 subaxial pedicles were included in the study after excluding those with vertebral artery filling defects and damage. Of the pedicles evaluated, 10 samples were from C3, 10 from C4, eight from C5, and two from C6. Intraobserver reliability demonstrated excellent agreement across all measurements (ICCs=0.87–0.96), indicating strong repeatability for SVA and SM, as well as other morphometric parameters.
Anatomical measurements
Table 1 presents the anatomical measurements and mean values for all variables. The average pedicle width (PW) was 6.08±1.22 mm, with the thickest pedicle seen at the C6 level (7.40±0.28 mm) and the thinnest at the C4 level (5.61±0.91 mm).
Mean values and statistical analysis results of anatomical measurements
The average VAA and TFA were computed as 27.97±8.09 mm2 and 5.09±2.00 mm2, respectively. The largest cross-sectional areas were seen at the C6 level (VAA=7.70±3.24 mm2; TFA: 32.50±9.19 mm2), whereas the smallest VAA and TFA were at the C4 (4.30±1.59 mm2) and C3 (22.50±9.09 mm2) levels, respectively. The average AOR was 18.80%±6.24%, with the C4 level showing the smallest AOR (15.40%±6.06%) and the C6 level with the largest value (23.50%±3.54%).
The mean values of coronal length of foramen (CLF) and coronal length of artery (CLA) were 6.47±0.86 mm and 2.49±0.43 mm, respectively, whereas the mean sagittal length of foramen (SLF) and sagittal length of artery (SLA) were measured as 5.27±0.79 mm and 2.43±0.53 mm, respectively.
The mean SM was calculated as 3.89±0.70 mm, with the widest SM at the C6 level (4.05±0.21 mm) and the narrowest at the C5 level (3.73±0.67 mm). When assessed separately, the average AFLD was 2.22±0.71 mm, and the average AFMD was 1.67±0.49 mm (Table 1).
ANOVA was used to assess whether the measured values for PW, TFA, VAA, APD, SM, CLF, SLF, CLA, and AOR varied between the cervical levels, yielding no statistically significant results. Only the SLA value was significantly different across cervical levels (p=0.033), with post hoc analysis revealing the difference being between C4 and C6 (p=0.04).
Descriptive statistics for the measured angle values are also shown in Table 1. The mean FMA and AMA values were 38.33°±4.87° and 33.00°±5.43°, respectively, whereas the average PA value was 46.43°±2.73°. Statistical analysis showed no significant difference among PA, FMA, and AMA values.
For the anatomical measurements (Table 2), positive correlations were observed for TFA with VAA (r=0.516), AFLD (r=0.579), and SM (r=0.664) (p<0.01 each). Additionally, a positive correlation was found between VAA and AOR (r=0.609, p<0.01), but no significant correlation was observed between AOR and TFA.
Correlation analysis of anatomical parameters
A negative correlation was found between the SM components AFMD and AFLD (r=−0.369, p<0.05), whereas a positive correlation was observed between SM and AFLD (r=0.753, p<0.01). A similar relationship was observed between AFMD and SM. AFMD showed a positive correlation only with APD (r=0.768, p<0.05).
Post-procedural findings
The average screw angle (SA) was found to be 27.17°±11.53°, whereas the average ASV was 3.55±3.13 mm. ANOVA testing showed no significant differences between cervical levels for SA (p=0.779) and ASV (p=0.464).
Macroscopic and microscopic examinations revealed that 11 (36%) of the 30 vertebral artery specimens adjacent to the screws had full-thickness vessel damage, and 10 (33%) showed fragmentation of the internal or external lamina fibers without vascular layer damage, whereas 9 (30%) had no damage.
Relationship between histologic damage and screw position
The cervical levels with injured vertebral artery were evaluated based on the number and angle of screws penetrating the foramen at the undamaged level. In the injured group, although screws were inserted at a narrower angle, there was no statistically significant difference in SA; however, ASV was significantly greater in the injured group compared to the uninjured specimens.
Likewise, vertebral artery injuries were analyzed by grouping them according to previously defined safe zones (Table 3). Two groups were formed based on whether ASV exceeded the APD [12,19]; statistical analysis revealed no significant correlation between these groups (p=1.00). Similarly, damage-analysis grouping based on whether SA was smaller than the AMA [21] revealed no significant correlation between the groups (p=0.199). Notably, another subgroup analysis based on whether ASV exceeded the SM revealed a statistically significant correlation (p<0.001) when cases with ASV greater than the SM were compared with those with vertebral damage [17]. Cramer’s V value was computed as 0.791, indicating a positive correlation between these variables (p<0.001).
Statistical analysis of damage status and parameters
Discussion
In most cases, cervical pedicle screws get displaced laterally, likely damaging the vertebral artery within the transverse foramen [11]. Although vertebral artery injury is rare, it remains one of the most devastating complications emanating from cervical spine surgery, with complication rates being less than 1% in anterior cervical procedures, to about 5% in posterior procedures.
Consistent with the findings of previous anatomical studies, in our specimens, the vertebral artery was predominantly located medially within the transverse foramen [22,23], which was evident from the greater mean AFLD than the mean AFMD noted in the analysis. Furthermore, the positive correlation between SM and AFLD suggests that the lack of correlation with AFMD may be because of the transverse foramen widening more laterally, as a result of which the vertebral artery remains medially located. The AFMD–APD correlation is likely because both parameters relate to the medial region of the vertebral artery.
Regarding anatomical measurements, we observed that TFA was positively correlated with VAA, AFLD, and SM, which is consistent with the results described by Kim et al. [22]. These findings support the idea that a wider transverse foramen contains a larger vertebral artery, thereby increasing the SM. Alternatively, it can be said that the increased SM in a wider transverse foramen is because of greater AFLD, suggesting that in a wider transverse foramen, the pedicle screw has a greater SM against damaging the vertebral artery, making the procedure safer.
Although cervical vertebrae show morphological changes based on anatomical levels, our study found no significant differences in key anatomical measurements across levels, except for the SLA. This is likely because the subaxial vertebrae have a more uniform structure compared to the atypical C1, C2, and C7 vertebrae.
Previous studies with larger samples have reported a 5%–24% rate of pedicle screws breaching the lateral wall, yet none reported clinical findings related to vertebral artery injury [13,20,24–28]. The lower-than-expected rate of vertebral artery–related complications in cases with cervical pedicle screw breaches into the transverse foramen has led to the concept of a “safe zone” among surgeons. In this study, we aimed to establish the practical relevance of this concept, which has been previously defined by various authors as different anatomical gaps and measurements. Tomasino et al. [12] identified the APD value as the critical threshold for vertebral artery injury, whereas Lee et al. [21] indicated the AMA value as the critical threshold. They suggested that screw malposition within these thresholds would not injure the artery as it remains within a safe zone; however, our study found no significant correlation between vertebral artery injury and exceedance of the AMA or APD threshold.
Based on anatomical measurements and CT studies, it has been suggested that the safe zone extends beyond the distance from the pedicle, encompassing the entire space within the transverse foramen, as a misplaced screw can mobilize the vertebral artery in any direction [17]. Our analysis was based on the safety margin described by Sharma et al. [17] and showed a significant correlation between screws violating the safety margin and the vertebral artery injury. This indicates that when pedicle screws breach the transverse foramen, the vertebral artery is not fixed in place but can move laterally without being damaged. Our study confirms that the space within the transverse foramen, excluding the vertebral artery, indeed serves as a safety margin. This was further confirmed in our pressurized cadaver model when screws that violated the safety margin (SVA > SM) resulted in arterial wall injury confirmed by histological analysis, whereas breaches within the SM did not produce structural damage to the vertebral artery. Thus, limited lateral screw breach may result in arterial displacement rather than true vessel disruption; alternatively, exceeding the SM is associated with a high risk of mechanical penetration and vascular injury.
Recent developments in cervical pedicle screw insertion techniques have focused on improving the lateral trajectory control to minimize the risk of vertebral artery compromise. Apart from technological solutions, such as robotic assistance and navigation systems, which enhance the accuracy and reduce insertional deviations, improvements in freehand techniques have strengthened anatomical reproducibility and safety in cervical spine procedures [29,30]. The medial pedicle pivot point technique introduced by Kwon et al. [31] represents a notable paradigm shift by using the medial pedicle cortex as a controlled fulcrum to optimize the screw trajectory. Although our cadaveric model did not evaluate different surgical techniques, our findings provide anatomical validation to complement these approaches by clarifying the quantifiable SM available within the transverse foramen before vertebral artery penetration occurs.
A critical implication of our study’s findings is the need to evaluate pedicle morphology with preoperative CT imaging before selecting screw size and trajectory. In our cadaveric specimens, the pedicle width was slightly larger than that reported by Kwon et al. [31], underscoring that pedicle dimensions vary substantially across individuals. When the screw diameter is close to the pedicle width, the residual cortical thickness becomes minimal, increasing the probability of a lateral wall breach or biomechanical compromise even with small deviations in trajectory.
We found a significant correlation between VAA and AOR, indicating that the dominant vertebral artery occupies more space within the transverse foramen compared to the nondominant or co-dominant artery. This correlation between VAA and SM suggests that larger vertebral arteries provide greater SM. Clinically, this means that injury to the dominant vertebral artery could result in more severe posterior circulation impairment, leading to pronounced clinical symptoms; conversely, injury to the co-dominant or nondominant vertebral artery might result in a relatively silent clinical presentation because of the compensation by the dominant vertebral artery in maintaining posterior circulation.
Although this study was conducted using a cadaveric model, its findings carry meaningful clinical relevance. The quantification of the SM within the transverse foramen provides a robust anatomical and practical explanation for why many lateral pedicle screw breaches reported in the literature do not result in vertebral artery injury or neurological symptoms. These findings suggest that when lateral violation remains within the SM range, the artery may be displaced rather than injured, allowing for safe observation in selected patients without immediate revision surgery, as long as there is no radiologic evidence of arterial compression or contrast extravasation. However, when the SM is exceeded or a vascular injury is suspected based on intraoperative findings or postoperative angiography, early endovascular or surgical intervention is crucial to prevent complications. Our results bridge the gap between cadaveric anatomy and clinical decision-making by supporting an individualized approach to revision surgery following lateral pedicle breach. Therefore, SM should not be interpreted as a guarantee of safety but rather as an anatomical guideline that must be corroborated with clinical findings, neurovascular imaging, and intraoperative judgment when considering revision surgery.
This study has several limitations. First, it was conducted on few cadaveric specimens, all of which were male, limiting the generalizability of our findings to female anatomy and broader populations. Although a pressurized perfusion system was used to simulate physiological intraluminal pressure, true arterial pulsations and vasoregulatory mechanisms could not be reproduced in this study. For this reason, a binary histologic classification (injury present/absent) was intentionally used to maintain objectivity and reproducibility as detailed grading of arterial injury in cadaveric tissue may not accurately reflect in vivo hemodynamic consequences. It is noteworthy that even in the absence of direct arterial penetration, prolonged mechanical compression or intimal irritation may predispose to the risk of delayed vascular complications, including thrombosis, dissection, or pseudoaneurysm formation [32,33]. Therefore, lateral pedicle breach within the SM should not be interpreted as completely risk-free and warrants careful clinical and radiological follow-up when encountered intraoperatively or on postoperative imaging. Finally, all screws were deliberately inserted to produce lateral violations without navigation or robotic guidance, which may not fully replicate real-world surgical conditions; however, this experimental design was necessary to evaluate the anatomical threshold of vertebral artery tolerance.
Conclusions
This cadaveric study employed a quantitative methodology to demonstrate that a measurable SM exists between the pedicle wall and the vertebral artery within the transverse foramen. Our findings reveal that not all lateral pedicle screw breaches carry the same risk, requiring revision surgery; this decision must be individualized based on breach severity, vascular imaging, and clinical symptoms rather than performed routinely. When the SM is not exceeded, the artery may be displaced without structural damage; however, when the SM is violated, the risk of direct arterial penetration is tangible, and early endovascular or surgical intervention may be warranted. While these findings provide an anatomical foundation for clinical decision-making, prospective clinical studies are needed to validate this anatomical threshold and assess its influence on surgical outcomes.
Key Points
This study aimed to determine the critical threshold at which subaxial cervical pedicle screw malpositioning can lead to vertebral artery injury.
The safety margin, defined as the combined distance from the vertebral artery to the lateral and medial walls of the transverse foramen, represents the critical region for vertebral artery injury.
Occasionally, the vessel may be displaced and not injured by the malpositioned screw; thus, revision surgery should be considered based on the patient’s clinical condition and breach status rather than the screw position alone.
As it is a cadaveric study, physiological blood flow and risk of ischemia could not be assessed.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Funding
This study was supported by Marmara University Scientific Research Projects Unit with project number 10949.
Author Contributions
Each author should declare their role in accordance with the CRediT Taxonomy initiative (https://credit.niso.org). Conceptualization: ÖS, YG. Data curation: ÖS. Formal analysis: ÖS. Methodology: ÖS, YG, BG, KK. Investigation: ÖS, YG, BG, KK. Project administration: ÖS. Visualization: ÖS. Writing–original draft: ÖS. Writing–review & editing: ÖS, YG, BG, KK. Supervision: YG. Final approval of the manuscript: all authors.