Introduction
Posterior atlantoaxial and occipitoaxial fixation are commonly employed in the treatment of craniovertebral junction disorders. Various internal fixation techniques have been developed, including the Gallie and Brooks posterior wiring procedures, interlaminar clamps, Magerl transarticular screws, and the Goel-Harms technique [1-4]. Among these, the Goel-Harms atlanto-axial fixation has gained widespread acceptance due to its superior biomechanical stability, particularly in cases of atlantoaxial instability or dislocation. This technique typically involves the individual fixation of the C1 lateral mass and C2 pedicle with screws, rods, or plates. Despite its popularity, a significant concern remains: the risk of vertebral artery injury during screw insertion, especially in patients with a high-riding vertebral artery and a markedly narrowed C2 pedicle [5-7].
The “in-out-in” screw fixation technique was first described for use in the thoracic spine [8]. This method allows anchoring of the screw in three or four cortical layers in addition to the cancellous bone of the vertebral body. The potential benefits of this “in-out-in” screw include enhanced safety of neurological structures and increased bone-screw interface stability in cases of severely narrowed thoracic spine pedicles.
In this report, we describe a transforaminal “in-out-in” screw technique for C2 fixation. This novel approach employs an entry point similar to that of the conventional C2 pedicle screw but is directed more caudally and medially toward the C2 vertebral body. The screw trajectory traverses the upper part of the intervertebral foramen, circumventing the vertebral artery laterally and the spinal cord medially. This transforaminal “in-out-in” screw may achieve rigid three-column fixation with three or four cortical purchases.
To our knowledge, the transforaminal “in-out-in” screw fixation technique has not been previously described. Therefore, we conducted a multiplanar computed tomography (CT)-based anatomical correlation study along with a preliminary clinical investigation to evaluate the feasibility of this multi-cortical purchase screw fixation technique.
Technical Notes
Patient demographics
Between October 2022 and March 2023, transforaminal “in-out-in” screws were used in nine patients (seven females, two males; mean age, 53.7 years; range, 15–70 years) who underwent posterior atlantoaxial internal fixation. Indications included traumatic atlantoaxial instability (n=1), atlantoaxial dislocation with os odontoideum (n=1), degenerative atlantoaxial dislocation (n=1), acute odontoid fractures (n=2), old odontoid fractures with atlantoaxial instability (n=2), and atlantoaxial dislocation due to atlas occipitalization with basilar invagination (n=2). All patients had severe hypoplasia of the C2 pedicle on at least one side, precluding conventional C2 pedicle screw insertion. Neutral alignment and flexion-extension lateral radiographs (except in patients with acute fractures), CT angiography, and magnetic resonance imaging (MRI) were obtained for all patients. Severe hypoplasia of the C2 pedicle was defined as a diameter less than 3.5 mm [9]. For each planned transforaminal “in-out-in” screw, meticulous preoperative radiological measurements were performed using multiplanar CT reconstruction images. The appropriate medial inclination angle and planned screw total length were measured on the axial plane, while the C2–3 intervertebral foraminal height was measured on the sagittal plane. The total length of the planned screw trajectory was divided into three segments corresponding to the three parts (“in-out-in”). The first segment (S1) represents the length of the screw within the bony part of the C2 inferior articular process; the second segment (S2) is the portion traversing the intervertebral foramen; and the third segment (S3) is the length of the screw within the bony part of the C2 vertebral body (Fig. 1). The lengths of all three segments were measured. Patient demographics and radiological measurements are summarized in Table 1. Informed consent was waived due to the retrospective nature of the study.
Surgical procedure
After general anesthesia, patients with atlantoaxial instability were positioned prone. Cervical alignment and stabilization were achieved using either Gardner-Wells traction or a Sugita head holder. A 30° head-up tilt was applied to provide countertraction and reduce venous congestion. The cervical spine was exposed subperiosteally from the occiput to the C3–C4 level. The C1–C2 complex was exposed to the lateral border of the C1–C2 articulation by retracting the paravertebral muscles. Bilateral C1 lateral mass screws were inserted either from the posterior surface of the C1 lateral mass, as described by Goel and Harms, or from the surface of the C1 posterior arch, as described by Tan et al. [10]. An occipital plate was used in cases of atlas occipitalization. Standard C2 pedicle screws were inserted into newly developed C2 pedicles. In cases of severely narrowed C2 pedicles, with or without a high-riding vertebral artery, the transforaminal “in-out-in” screw technique was employed.
Prior to inserting the transforaminal “in-out-in” screw, the medial border and superior surface of the C2 isthmus were delineated using a Penfield dissector (No. 4). The screw entry point was identified and marked with a high-speed spur, typically located at the midpoint of the posterior surface of the C2 inferior articular process—similar to the entry point for a conventional C2 pedicle screw. A blunt-tipped pedicle awl was then advanced at an approximate 30° medial convergence angle, as determined by preoperative radiological measurements. Cephalad screw trajectory was confirmed using intraoperative lateral fluoroscopy. Once the blunt-tipped pedicle awl encountered the anterior cortex of the C2 inferior articular process, a 2-mm high-speed diamond burr was used to carefully perforate the anterior cortex of the C2 inferior articular process. The blunt-tipped pedicle awl was then reintroduced and advanced through the upper part of the intervertebral foramen until it contacted the posterolateral corner of the C2 vertebral body. At this point, a sharp-tipped pedicle awl was used to penetrate the cortex of the posterolateral corner and advanced anteromedially toward the anterior cortex of the C2 vertebral body. Intraoperative lateral fluoroscopy was repeated to confirm the accuracy of the pilot hole trajectory. A ball-tipped probe was then used to check the bony walls and measure the length of the pilot tract. A 3.5-mm polyaxial screw of appropriate length was inserted along the prepared pathway. In this cohort, all transforaminal screws used were 30 mm in length. A simplified schematic illustration depicting the screw trajectory in relation to the vertebral artery, spinal cord, and intervertebral foramen is presented in Fig. 2.
In patients with atlantoaxial dislocation, the surgical strategy followed the approach described by Wang et al. [11]. The patient was initially placed in the supine position, and skull traction equivalent to one-sixth of the body weight was applied using skull tongs. If skull traction resulted in less than a 50% reduction, a transoral atlantoaxial release was performed (one case in this series). If more than a 50% reduction was achieved, the patient was repositioned prone, and the posterior surgical procedure was carried out as described for patients with atlantoaxial instability.
In cases of atlantoaxial dislocation, reduction of the C1–C2 complex was achieved by applying an anteroinferior force to the C2 spinal process while simultaneously shortening the connecting rods. For patients requiring a definitive fusion, the posterior surfaces of the C1/occiput and C2 were decorticated, and cancellous bone harvested from the posterior iliac crest was placed over the decorticated areas to promote fusion. All patients were mobilized on the first postoperative day and instructed to wear a soft cervical collar for 3 weeks. Routine CT and MRI scans were performed to assess implant positioning and the extent of reduction. At the final follow-up, CT scans with multiplanar reconstructions were obtained to evaluate the status of the bony bridge in the fusion cases (Figs. 3–5).
Eleven transforaminal “in-out-in” screws were inserted in nine patients, all measuring 3.5 mm in diameter and 30 mm in length. One planned transforaminal “in-out-in” screw was aborted due to cerebrospinal fluid leakage and suspected dural sac violation during screw canal preparation; a C2 laminar screw was used as an alternative. The other eight screws were placed into normally developed C2 pedicles. No spinal cord injury, vascular injury, or other major complications were observed. Two patients experienced retroauricular neuralgia on the side of the transforaminal screw immediately after surgery; both cases resolved completely within 1 month. The mean follow-up duration was 19.6±1.3 months (range, 18.1–21.2 months). No implant failure was noted at the final follow-up.
Discussion
Since its introduction by Goel and Laheri [4] and Harms and Melcher [2], the combination of C1 lateral mass and C2 pedicle screws has become the standard atlantoaxial fixation pattern due to its superior biomechanical stability. However, the risk of vertebral artery injury remains a significant concern, particularly during C2 pedicle screw placement. In a study by Yeom et al. [7], approximately 21% of C2 pedicle screws resulted in violation of the vertebral artery groove, even in anatomically normal pedicles. The risk increases substantially in patients with high-riding vertebral arteries and severely narrowed C2 pedicles. Compounding this issue, the high-riding vertebral artery is often the dominant or even solitary vessel supplying the posterior brainstem [6,12]. Injury to such a vessel can lead to catastrophic consequences [13].
Given the extremely high risk of vertebral artery injury, several alternative techniques have been proposed as potentially safer substitutes for C2 pedicle screws in craniocervical and atlantoaxial fixation. One such option is the C2 laminar screw, which entirely avoids the vertebral artery trajectory [14]. However, it offers reduced biomechanical stability compared to C2 pedicle screws, which can achieve three-column fixation. Additionally, laminar screws pose a risk of anterior cortical breach and potential dural or spinal cord injury, particularly because the screw trajectory cannot be visualized intra-operatively [15]. In fusion procedures, the screw tips may also interfere with the preparation of the bone graft bed. Another alternative is the C2 isthmus screw, which terminates at the posterior wall of the vertebral artery groove to minimize the risk of vertebral artery injury [16]. However, the isthmus screw is biomechanically inferior due to its significantly shorter length compared to conventional C2 pedicle screws [17].
The so-called “in-out-in” screw was first used for the thoracic spine [8]. As an extrapedicular approach, it offers improved safety over traditional pedicle screws, particularly in the setting of severely narrowed thoracic pedicles, by reducing the risk of neural canal and spinal cord violation. Despite avoiding the pedicle, it can still achieve robust fixation by engaging up to three or four cortical layers and providing three-column fixation. More recently, the “in-out-in” screw technique has been adapted for C2 fixation. Guo et al. [18] described a C2 “in-out-in” screw fixation technique for atlantoaxial dislocation or instability. However, in their description, the screw trajectory violated the vertebral artery groove inferolaterally, such that the “out” part of the screw was in the vertebral artery groove of C2. To mitigate the risk of vascular injury, they used a Penfield dissector, extending from the medial border of the C2 transverse foramen to the roof of the vertebral artery groove, mobilizing the high-riding vertebral artery inferiorly by more than 3 mm, allowing safer surgical manipulation. Nonetheless, concerns persist regarding vertebral artery injury, and this technique is contraindicated in patients with a single dominant or solitary high-riding vertebral artery. Goel et al. [19] described a similar technique but with a more aggressive approach. They exposed and mobilized the vertebral artery by unroofing the vertebral artery groove in the C2 vertebral body. While this allows for direct visualization and manipulation of the artery, the procedure is technically demanding and carries a risk of massive bleeding during exposure of the vertebral artery.
Another variation of the C2 “in-out-in” screw technique was described by Du et al. [20] and Lee et al. [21]. In their approach, the inner cortex of the narrow C2 isthmus was drilled to create sufficient space for screw insertion while preserving the lateral cortex. The screw was directed along this drilled inner cortex into the anterior vertebral body, thereby achieving a three-column fixation with multi-cortical purchase. In this technique, the trajectory remained medial to the foramen transversarium to avoid vertebral artery violation, but at the cost of encroaching on the spinal canal. To protect the dura during drilling, the dural sac was slightly retracted medially by an assistant using a Penfield dissector. While this approach effectively reduces the risk of vertebral artery injury, it raises concerns regarding potential spinal cord injury due to dural sac manipulation and occupation of the spinal canal by the screw.
In this report, we described a modified C2 “in-out-in” screw fixation technique that is fundamentally different from all previously reported C2 “in-out-in” methods. According to anatomical studies, the average height of the C2–C3 intervertebral foramen is approximately 8.1 mm [22,23]. The dorsal and ventral C3 nerve roots lie at the bottom of the foramen, while the upper portion of the foramen is filled with adipose tissue [24,25]. These anatomical features make it feasible for a 3.5 mm diameter screw to be safely accommodated in the upper portion of the foramen without compromising neural structures. Based on this understanding, we developed our transforaminal “in-out-in” screw technique. The transforaminal “in-out-in” screw trajectory can be divided into three parts corresponding to the three components of its name. The first “in” refers to the screw’s initial entry into the bony part of the C2 inferior articular process; the “out” denotes its passage out the bony part, transversing the upper portion of the foramen; and the second “in” signifies the entry of the screw into the bony part of the C2 vertebral body. The critical aspect of this technique is ensuring that the foraminal part of the trajectory remains close to the inferior wall of the C2 isthmus, between the vertebral artery laterally and the dural sac medially. This requires meticulous preoperative planning using multiplanar CT reconstructions, as well as continuous intraoperative fluoroscopic confirmation. The screw sequentially passes through the inferior articular process, the intervertebral foramen, and finally into the vertebral body, with each segment corresponding precisely to the preoperative measurements. Another key aspect of this technique is the use of a blunt-tipped pedicle awl to probe the posterior lateral corner of the C2 vertebral body. Unlike the sharp-tipped pedicle awl or high-speed burr, the blunt end significantly reduces the risk of inadvertent injury to the vertebral artery or dural sac as it navigates the intervertebral foramen. In our cohort, one planned transforaminal “in-out-in” screw placement was suspended due to suspected dural sac injury during canal preparation. This incident likely resulted from using a sharp-tipped awl and directing it too medially out of concern for vertebral artery violation laterally. Following this event, we adopted the blunt-tipped awl for probing the posterolateral corner of the vertebral body, and no further dural sac injury or cerebral fluid leakage occurred during the operation.
The trajectory of the transforaminal “in-out-in” screw significantly reduces the risk of vertebral artery and spinal cord injuries. This three-column fixation provides superior biomechanical strength with the purchase of three or even four cortices. Preliminary clinical results were satisfactory: no vertebral artery or spinal cord injuries were observed, and no cases of implant failure occurred during follow-up. In patients requiring fusion, bony bridging was confirmed on CT reconstruction images. Two patients reported retroauricular neuralgia on the side of the transforaminal screw immediately after surgery, possibly due to C3 nerve root irritation [26]. Both patients experienced complete resolution of symptoms within 1 month.
This study has some limitations. First, it represents a preliminary technical report with a small sample size and a relatively short follow-up period. Larger prospective clinical studies with longer follow-up are warranted. Second, we did not perform cadaveric biomechanical study to objectively assess the strength of this modified “in-out-in” screw. Although the three-column fixation with multi-cortical purchase is theoretically robust, future in vitro biomechanical studies—such as pull-out strength testing—are needed to provide more convincing evidence of its mechanical superiority compared to alternative fixation methods, including C2 laminar screws or short C2 isthmus screws.
In conclusion, the transforaminal “in-out-in” screw may achieve rigid three-column fixation with three or four cortical purchases. The screw trajectory avoids both the vertebral artery laterally and spinal cord medially, while rarely disturbing the exiting nerve root during the passage through the upper part of the intervertebral foramen. It provides a safe and effective alternative for fixation when the traditional pedicle screw placement is not feasible due to a severely narrowed C2 pedicle.
Key Points
The anatomic structure of the C2–3 intervertebral foramen can accommodate a 3.5 mm diameter screw.
The trajectory of the transforaminal “in-out-in” screw avoids both the vertebral artery laterally and the spinal cord medially at the upper portion of the C2–3 intervertebral foramen.
This transforaminal “in-out-in” screw may provide rigid three-column fixation by achieving three or even four cortical purchases.
Safe placement of the transforaminal “in-out-in” screw is feasible with meticulous preoperative radiological measurements and intraoperative fluoroscopic confirmation.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
