Risk of neurologic deficit in medially breached pedicle screws assessed by computed tomography: a systematic review
Article information
Abstract
Pedicle screws are commonly used for vertebral instrumentation, and a postoperative computed tomography (CT) scan is used to evaluate their position within the pedicle. Medial pedicle screw breaching occurs in 20%–40% of cases. This study investigated the correlation between radiographically evident medial breaching and the incidence of nerve injury, shedding light on the clinical implications. A literature search was conducted on biomedical databases regarding neurologic deficits associated with medially breached pedicle screws with pre-defined inclusion and exclusion criteria. The methodology of the included studies was analyzed, and a systematic review and meta-analysis were performed to investigate the correlation between medial breach on axial CT and clinical neurologic deficits. Our study included thirteen articles. Medial breaches <2 mm caused no neurologic deficit. Medial breaches of 2–4 mm increased the risk of neurologic deficit by 83%, with a risk ratio of 0.17. Breaches exceeding 4 mm increased the risk by 90%, with a risk ratio of 0.1, and were associated with radiculopathy or muscle weakness in 25%–100% of cases. Medial pedicle screw breaches <2 mm are safe, carrying no risk of neurologic injury. Breaches exceeding ≥2 mm significantly increase this risk. For patients experiencing new neurologic deficit (sensory or motor) after pedicle screw instrumentation, particularly in lumbar vertebrae, a postoperative axial CT scan is recommended to identify breaches exceeding 2 mm as the potential cause of neurologic deficit.
Introduction
Pedicle screw insertion is technically demanding due to the complex neurovascular anatomy around the spinal canal. Pedicle screw is used more frequently in lumbar vertebrae compared to other vertebral segments. This is because the thicker pedicles of lumbar vertebrae facilitate cannulation, and their trajectories minimize the risk of compromising important neurovascular structures. This reduces the likelihood of serious neural damage due to medial breach of the screw because the components of cauda equina are much less prone to damage. In contrast, medial breach of pedicle screws in non-lumbar vertebrae poses a significant risk of spinal cord injury [1].
Several methods are employed for pedicle screw placement, including robot-assisted, computed tomography (CT) navigation-guided, fluoroscopy-assisted, and free-hand techniques. A meta-analysis found no significant difference in pedicle screw placement accuracy among these four techniques. The accuracy rates for robot-assisted technique, CT navigation-guided technique, fluoroscopic-assisted technique, and free-hand technique were 90.5%, 95.5%, 91.5%, and 93.1%, respectively [2]. Due to proximity to the spinal canal, pedicle screw breaches, particularly medial breaches, can lead to disastrous complications. Therefore, accurate and safe placement of the screw within the pedicle is a crucial step during surgery.
Several in vitro and in vivo studies have evaluated the effect of pedicle screw breaches on neurological status. Medial pedicle screw breaches, regardless of associated neurologic injury, occur in a significant proportion of cases, ranging from 20% to 40%. This incidence rate appears to be technique-independent, unaffected by the specific method used for pedicle screw instrumentation [3]. A previous study showed false-negative findings for pedicle screw breaching, with 0.3% of the screws causing no intraoperative neurologic deficits despite postoperative CT confirming medial pedicle screw breaches. This suggests that not all medial pedicle screw breaches cause neurologic deficit. However, due to the proximity of the medial side of the pedicles to important structures, such as the spinal canal and neural elements, neurologic deficits remain the most common clinical manifestation of medial pedicle screw breaches [4].
Surgeons commonly use postoperative CT scans to assess pedicle screw accuracy and detect potential breaches, particularly medial breaches. Although these concerns are warranted, CT scan findings do not always accurately represent pedicle screw breaching. Therefore, this can lead to unnecessary surgical delays, additional examination charges, and even medico-legal disputes [5].
The acceptable degree of medial pedicle breaching remains a topic of debate. Gertzbein and Robbins [6] suggested a 4-mm safe zone as an allowable limit for medial breaching. However, Belmont et al. [7] proposed a more conservative threshold of 2 mm in the thoracic spine. Castro et al. [8] found that all patients with postoperative neurologic deficits had medial pedicle screw breaches of ≥6 mm. Patients with spinal deformity present additional challenges, as the thecal sac is often displaced toward the concave side of the spine, complicating instrumentation on this side.
This systematic review investigated the relationship between radiologically confirmed medial pedicle screw breaching and the incidence of nerve injury, aiming to clarify whether all observed medial breaches can cause spinal cord injury. The specific objectives were (1) to investigate the correlation between radiographically evident medial breaching and the incidence of nerve injury and (2) to develop a risk stratification model for midline crossing or medial breach of pedicle screws, enhancing our understanding of the clinical implications of these radiographic findings.
Materials and Methods
This systematic review adheres to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines [9]. The review aimed to assess the association between medial pedicle screw breaches observed on axial CT and clinical neurologic deficit. Using the PICO framework (patient/population/problem, comparison, intervention, and outcome), the following research questions were formulated: (1) problem: medially pedicle screw breach; (2) intervention: postoperative axial CT; (3) comparison: pedicle screw in perfect position; and (4) outcome: neurologic deficit (sensory or motor).
A literature search was conducted in the Cochrane Library, Medline/PubMed, and EBSCOhost databases using keywords, index terms (including MeSH/Medical Subject Headings), and Boolean operators. All three databases were searched up to June 2024. Bibliographies of the included articles were also screened to identify additional studies.
The inclusion criteria were as follows: (1) type of study: randomized controlled trials (RCTs) or non-RCTs evaluating neurologic injury or deficit after pedicle screw medial breaching; (2) types of participants: patients with spinal diseases (trauma, deformity, degenerative, or infection) that require pedicle screw insertion in thoracal or lumbar vertebrae; (3) intervention: spinal diseases that need pedicle screw placement without certain indication; and (4) outcomes evaluation: clinical neurologic deficits including radiculopathy or myelopathy.
The exclusion criteria were as follows: (1) cohort studies, review articles, meta-analyses, case reports or case series, and expert opinion; (2) animal studies; and (3) articles published in languages other than English.
The methodological quality of each study included in the review was assessed using the level of evidence in medicine criteria [10]. Studies were also assessed for risk of bias using the criteria presented in the Cochrane Handbook. Assessment of quality and risk of bias in RCTs was conducted using domain-based evaluation while the Newcastle-Ottawa Scale was used for non-RCTs, as recommended by the Collaboration [11]. Meta-analysis was performed to assess the risk ratio of medial breach, with 95% confidence intervals.
Results
Literature search
A schematic illustration of the literature search is presented in Fig. 1. After applying inclusion and exclusion criteria, 13 articles were included in the review. Database search retrieved a total of 135 articles (82 from PubMed/Medline, 11 from EBSCOhost, and 44 from Cochrane Library databases), with 11 articles retrieved from cross references. Fifty-seven articles were excluded after a review of titles, 43 were excluded after reviewing the abstracts, and 29 articles were excluded after full-text reviews. Two articles were excluded due to duplication and two articles each were excluded due to the non-availability of full-text and non-recruitment status of studies, respectively. After reviewing the cross references, four articles each were excluded after review of the abstract and full-text review, respectively.
Chart of the literature search.
The systematic review included five RCTs and eight non-RCTs. The research methods used in RCTs were assessed using domain-based evaluation. The methodological assessment revealed that all studies were susceptible to bias from various sources, such as sample size and outcome assessment timing. The results of the methodology assessment for RCTs are presented in Fig. 2 [12–16]. Quality assessment of non-RCTs was performed using the Newcastle-Ottawa Scale, showing that all were of good quality (Table 1) [17–24]. A summary of the included studies is presented in Table 2 [12–24].
Assessment of the quality of randomized controlled trials study using domain-based evaluation.
Quality assessment of the non-randomized controlled trials by Newcastle-Ottawa Scale
Summary of the included studies
Risk of neurologic deficit in medially breached pedicle screws
Fig. 3 presents the funnel plot assessment for publication bias. The plot’s asymmetric shape suggests the presence of publication bias, indicating that some studies with negative results may be missing from the literature.
Funnel plot. SE, standard error; RD, risk of difference.
Thoracic vertebrae
As shown in Table 2, only three studies evaluated medial breach in thoracic vertebrae. Chen at al. [18] found six and three cases of <2 mm and 2–4 mm medial breach, respectively. However, no neurologic deficits were encountered. Nevzati et al. [21] found no medial breach in all cases of thoracic pedicle screw instrumentation.
Lumbar vertebrae
Thirteen studies had evaluated medial breaches in lumbar vertebrae. None of the <2 mm medial breaches in lumbar vertebrae caused any neurologic deficit. Only one case with 24 mm medial breach developed radiculopathy, requiring surgical revision. For breaches exceeding 4 mm, the risk of neurologic deficits—manifesting as radiculopathy, numbness, or muscle weakness—ranged from 25% to 100%. Seven studies, including those by Qian et al. [17], Chen et al. [18], Chang et al. [25], Ohmori et al. [12], Nayar et al. [13], Kim et al. [15], and Malham Munday [16], did not report any medial breaches exceeding 4 mm. The incidence of neurologic deficits due to medial pedicle breach varied across studies. Ringel et al. [14] reported deficits in 25% of cases, Yoshida et al. [19] in 30% of cases, Saarenpaa et al. [20] in 56.25% of cases, Nevzati et al. [21] in 75% of cases, and Spitz et al. [22] and Arnaout et al. [23] in 100% of cases. Only Kim et al. [15] found a neurologic deficit in a medial breach of 2–4 mm.
Meta-analysis was performed for pedicle screw breaches in lumbar vertebrae. No neurologic injuries occurred in cases without pedicle breach or medial pedicle screw breaches <2 mm. Due to the absence of events in both groups, risk ratio estimation and heterogeneity assessment were not applicable (Fig. 4) [12–24]. Only one case of neurologic deficit was reported in a medial pedicle screw breach of 2–4 mm. The risk ratio for neurologic deficit in this category was 0.17 (Fig. 5) [12–24]. The heterogeneity assessment was not applicable. For medial breaches >4 mm, the risk ratio for neurologic deficit was 0.1 (Fig. 6) [14,17,19–23], with minimal heterogeneity (I2=0%).
Risk of neurologic deficit in medial pedicle screw breach of less than 2 mm. M–H, Mantel-Haenszel; CI, confidence interval.
Risk of neurologic deficit in medial pedicle screw breach of 2–4 mm. M–H, Mantel-Haenszel; CI, confidence interval.
Risk of neurologic deficit in medial pedicle screw breach of >4 mm. M–H, Mantel-Haenszel; CI, confidence interval.
Discussion
To our knowledge, this is the first systematic review assessing the risks accruing from medial breach of pedicle screw. Most existing grading systems measure breach severity in millimeters from the outer cortex surface of the medial pedicle to the outer side of the breached screw. We employed the classification system by Gertzbein and Robbins [6], which categorizes pedicle screw breaches based on the extent of the medial breach. This classification system defines breaches as grade A when the screw is within the pedicle, grade B for breaches less than 2 mm, grade C for breaches between 2 and 4 mm, and grade D for breaches exceeding 4 mm. All of the included studies had used this classification. Yoshida et al. [19] and Spitz et al. [22] classified pedicle screw breaches into four grades: grade A (screw within the pedicle); grade B (medial breach <2 mm); grade C (medial breach 2–4 mm), grade D (medial breach >4 mm). Arnaout et al. [23] also employed a four-grade classification: adequate (screw within the pedicle); minor (medial breach <2 mm); moderate (medial breach 2–4 mm), and severe (medial breach >5 mm).
In the thoracic vertebrae, the distance between the screw and the adjacent nerve roots is relatively narrow, ranging inferiorly from 1.7 to 2.8 mm and superiorly from 1.9 to 3.9 mm [26]. However, our review found that medial breaches of thoracic vertebrae (both <2 mm and 2–4 mm) did not cause neurologic deficits; thus, we did not continue the meta-analysis. Despite less tolerance to medial pedicle screw breach in the thoracic spine due to the absence of epidural space between the dural sac and pedicle, we found no neurologic deficit in thoracic breaches.
According to literature, there are two possible causes of frequent pedicle screw breaches in lumbar vertebrae. First, certain spinal pathologies, particularly ankylosing spondylitis, can lead to chronic fibrosis and increased muscle tension, making it more challenging to expose the posterior elements. Second, the routine use of 7.5 mm screw diameters in L3–S1 vertebral segments, particularly in porotic bone, aims to enhance fixation strength and resist pull-out strength, potentially increasing the risk of breach [17].
In our review, none of the <2 mm medial breaches in lumbar vertebrae caused any neurologic deficit. Only one case of 2–4 mm medial breach resulted in radiculopathy, necessitating surgical revision. Our meta-analysis revealed that medial breaches of 2–4 mm increase the risk of neurologic deficit by 83%, with a risk ratio of 0.17. For breaches exceeding 4 mm, the risk of neurologic deficit increases by 90%, with a risk ratio of 0.1. Breaches <2 mm pose no risk of neurologic injury, whereas breaches of 2 mm or more significantly increase the risk. The heterogeneity assessment was only applicable to medial breach of >4 mm, with the I2 value of 0%, indicating low heterogeneity.
Polly et al. [27] conducted a study of volumetric intrusion of pedicle screws in thoracic vertebrae, concluding that medial breaches of up to 4 mm do not typically result in neurologic compromise. Additionally, they suggested that acceptable pedicle screw placement is achieved when the medial breach is ≤2 mm. Macke et al. [28] evaluated the accuracy of robot-assisted pedicle screw insertion and found that 7.2% of the pedicle screws experienced medial breaches exceeding 2 mm. None of the medial pedicle breaches less than 2 mm resulted in neurologic deficit. The authors concluded that any medial breach greater than 2 mm is clinically significant. Spitz et al. [22] concluded that any breach greater than 2 mm is a significant breach. In their study, one patient required reoperation for screw repositioning due to a postoperative L-5 radiculopathy secondary to a grade D medial breach at L-5. This patient experienced improvement of the radiculopathy after reoperation.
While some studies suggest that medial pedicle screw breaches less than 4 mm do not cause neurologic deficit, Kim et al. [29] consider any breach exceeding 2 mm as unacceptable. Gertzbein and Robbins [6] identified a safe zone for medial breaches in lumbar vertebrae, up to 4 mm, without associated neurological deficits. They suggested that the initial 2 mm of breach falls within the epidural space, and the subsequent 2 mm lies within the subarachnoidal space [6]. Librianto et al. [1] also found that low-grade pedicle screw breaches did not cause neurologic deficits. Their study revealed higher risks of medial breach in the middle thoracic segment compared to other thoracic segments. The risk of medial breach was three times greater on the convex side of the deformity. Mac-Thiong et al. [30] emphasized the potential consequences of misplaced pedicle screw insertion, leading to neurologic deficits that may manifest early or late. Therefore, they strongly recommended removing the medially misplaced or breached screw, regardless of breach severity, due to the risk of early or late neurologic deficits. In addition, to neurologic deficits, a medially breached screw can cause headache due to orthostatic hypotension, as described in a case report by Albayram et al. [31]. The orthostatic headache in their patient was attributed to cerebrospinal fluid (CSF) leak or meningeal irritation in the absence of CSF due to medially breached screw. Orief et al. [32] found that new neurologic deficits can occur after severe medial pedicle screw breaches. This finding is consistent with the study by Gautschi et al. [33], which also demonstrated that neurologic deficits can result from misplaced screws.
Saarenpaa et al. [20] evaluated pedicle screw placement using axial CT. They found that the mean accuracy of pedicle screws remaining inside the pedicle was 85.7%, with 3.3% of the pedicle screws experiencing perforations of 4 mm or more outside the pedicles in medial and inferior directions. Notably, only 16.3% of the breached screws caused a neurologic deficit. Nevzati et al. [21] assessed the accuracy of free-hand pedicle screw insertion using axial CT. Their findings showed that 20% of the pedicle screws breached the pedicle. The breaches were categorized based on severity, with 0.9% being minor violations of ≤2 mm, 5.3% being moderate violations of 2.1–4 mm, and 3.8% being severe breaches exceeding 4 mm. Notably, 5.9% of the patients developed postoperative radiculopathy, and all of these patients had severe breaches (>4 mm). Arnaout et al. [23] recommended that radiologists consider any breach exceeding 4 mm as a red flag, warranting immediate notification to the surgeon to consider a revision plan.
Based on the results of this study, we recommend that patients with new neurologic deficits (sensory or motor) after pedicle screw instrumentation, particularly in lumbar vertebrae, undergo postoperative axial CT to evaluate pedicle screw breach, focusing on breaches exceeding 2 mm. If a breached pedicle screw is identified and correlates with clinical dermatome or myotome, surgical revision of the culprit screw should be considered.
To the best of our knowledge, this is the first systematic review to evaluate the risk of medial pedicle screw breach. However, some limitations should be acknowledged. The analysis did not account for screw size, specifically diameter, which may influence breach risks. Additionally, the underlying vertebral pathology necessitating pedicle screw instrumentation was not considered, despite potential variations in neurologic deficits across different spine pathologies.
Conclusions
Medial pedicle screw breaches of less than 2 mm are considered safe, posing no risk of neurologic injury. However, breaches exceeding 2 mm significantly increase this risk. For patients experiencing new neurologic deficits (sensory or motor) after pedicle screw instrumentation, particularly in the lumbar vertebrae, postoperative axial CT scans are recommended to identify breaches greater than 2 mm as the potential cause of neurologic deficit.
Key Points
In patients with new neurologic deficit (either sensory or motoric) after pedicle screw instrumentation particularly in lumbar vertebrae, postoperative axial computed tomography should be performed to evaluate pedicle screw breach of >2 mm as the culprit of neurologic deficit.
The medial breach pedicle screws <2 mm caused no neurologic deficit.
The risk of neurologic deficit in medial breach of 2–4 mm is increased by 83%.
The risk of neurologic deficit in >4 mm medial breach is increased by 90%.
Acknowledgments
Authors would like to say gratitude for Faculty of Medicine Universitas Indonesia and Dr. Cipto Mangunkusumo National Central Public Hospital for the opportunity to perform the study.
Notes
Author Contributions
Conceptualization: RM, WDH. Data curation: RM, WDH. Formal analysis: RM, WDH. Funding acquisition: RM, WDH. Investigation: RM, WDH. Methodology: RM, WDH. Project administration: RM, WDH. Software: RM, WDH. Resources: RM, WDH. Supervision: RM, WDH. Validation: RM, WDH, KDS. Visualization: RM, WDH, KDS. Writing–original draft: RM, WDH, KDS. Writing–review and editing: RM, WDH, KDS, DFK. Final approval of the manuscript: all authors.