C1-C2 transarticular screw fixation for atlantoaxial instability.

Objectives: The atlantoaxial segment of the cervical spine is commonly destabilized in a variety of disorders. Transarticular screw fixation of the C1-C2 joint has been proposed as a biomechanically superior therapeutic modality. The authors present their experience with this technique.

Methods: A retrospective analysis of 23 patients treated with this technique was performed. The mean follow-up period was 39.5 [+ or -] 0.1 months.

Results: Mean duration of hospitalization was 3.4 [+ or -] 0.1 days (range, 2 to 11 days). No intraoperative or early postoperative complications were detected. Four patients (17.4%) had postoperative complications unrelated to the primary procedure. The position of the screw was judged as satisfactory in 21 patients (91.3%). Two patients (8.7%) with suboptimal positioning of the screws were neurologically intact but needed no reoperation. Solid osseous fusion was detected in 19 patients (82.6%).

Conclusions: Transarticular C1-C2 screw fixation appears to be a safe and surgically reliable technique. Criteria for its application and refinements in its technical considerations continue to advance its clinically versatile therapeutic potential.

Key Words: atlantoaxial, fusion, instability, screw, transarticular

**********

It is well known that the atlantoaxial segment is the most mobile region of the entire vertebral column. Approximately 50% of the rotation of the cervical spine occurs at the C1-C2 joint. (1) The average rotation at the C1-C2 joint has been calculated to be 43 degrees (range, 32 to 50 degrees). (2) When destabilized, its range of motion increases significantly, with consequent risk of severe neuronal damage. The multifactorial causality of the C1-C2 instability may be secondary to various pathologic conditions such as trauma (odontoid fracture, Jefferson fracture, traumatic spondylolisthesis of axis, ligamentous injury), congenital anomalies, skeletal dysplasia, Down syndrome, rheumatoid arthritis, osteoarthritis, infections, or tumors.

Several posterior surgical approaches have been used to achieve atlantoaxial arthrodesis, such as the Gallie technique (1939), (3) Brooks and Jenkins technique (1978), (4) or interlaminal clamping (1982). (5) In 1979, Magerl and Seemann et al (6) reported the results of a posterior C1-C2 transarticular screw fixation technique (TAS). This approach is particularly indicated in patients with conditions that are refractory to conventional wiring techniques, patients with nonunion or delayed union of odontoid fractures with atrophic changes at the fracture site, patients with an associated C1 fracture, patients in whom laminectomy is required, patients with extreme atlantoaxial instability secondary to os odontoideum or rheumatoid arthritis, and patients in whom external immobilization is difficult or contraindicated. (6) The TAS technique has reportedly had higher fusion rates (87 to 100%) compared with other techniques, and previously published biomechanical studies have proven that this surgical approach provides the rigidity and rotational stability at the fusion site needed to maximize fusion success. (6-12) In addition, this method obviates the need for rigid external bracing after surgery, as the biomechanical strength of the fixated atlantoaxial complex is excellent. (9)

The major drawback of using the TAS approach is associated with the degree of technical complexity of the procedure itself, and its steep learning curve. The risk of potential injury to the vertebral arteries, the spinal cord, the underlying dura, and adjacent exiting nerve roots is prominent because of the close proximity of the screw path to these important structures. Furthermore, in combination with the variable size and location of the transverse foramen of the lateral mass of both the atlas and axis, abnormalities such as comminuted fractures of C1 or C2, anomalous course of the vertebral artery, or existence of an aberrant vertebral artery, make the optimal insertion of the C1-C2 transarticular screws quite demanding. (13)

The already long list of complications associated with this procedure keeps expanding. It includes malpositioning of screws, long-term implant failure, dural tear, neurologic deficit (including suboccipital paresthesia or hypoglossal paresis), and vertebral artery injuries, ranging from occult arteriovenous fistula to frank arterial compromise causing brainstem infarction and death.

Meticulous knowledge of the regional anatomy, detailed preoperative examination of the C1-C2 complex with computed tomography (CT), and careful insertion of screws under fluoroscopic guidance are essential elements for successful placement of transarticular screws.

In this retrospective clinical study, we report our experience with posterior atlantoaxial transarticular screw fixation and fusion with cannulated screws. We also discuss our observations and thoughts regarding a few tricky technical points of this procedure.

Materials and Methods

Over a period of 5 years (January 1, 1996, to December 31, 2000), 48 patients were admitted to our institution with the diagnosis of atlantoaxial instability. Seventeen patients were treated with posterior wiring stabilization techniques (managing surgeon's decision), whereas in 31 patients, transarticular screw fixation was considered as the first treatment option by the involved neurosurgeon. In six of these patients, there was evidence of severe osteoporosis, and they were treated conservatively with prolonged external immobilization. In two other patients of this group, the preoperative imaging evaluation revealed an abnormal course of the vertebral arteries, and the patients were finally treated with a posterior wiring approach. Atlantoaxial screw fixation was performed in 23 patients (17 males and 6 females) with atlantoaxial instability. Their ages ranged from 19 to 81 years (mean age, 46.3 [+ or -] 0.2 years). In regard to the causality of their atlantoaxial instability, 16 patients had posttraumatic instability, 4 patients had instability secondary to advanced rheumatoid arthritis, and 3 patients had os odontoideum. Among these patients, 17 were treated at our institution from the beginning, whereas 6 patients were initially treated at another institution using various posterior wiring fusion techniques, with no success.

The prominent presenting symptoms in our series were as follows: in 18 patients, moderate to severe occipital pain; in 3 patients, neck pain radiating to the shoulders; and in 2 patients, symptomatology of progressive myelopathy (positive Lhermitte sign). All of our patients complained of some neck tenderness or pain on admission.

The diagnosis of atlantoaxial instability was made on the basis of complete cervical spine plain radiographs. The radiographic evaluation included fine (1-mm slice thickness) CT scan of the cervical spine with coronal, sagittal, and special reconstructions to delineate the bony anatomy of the upper cervical spine and to visualize the potential pathway of a transarticular screw. Magnetic resonance imaging was routinely used for evaluating the spinal cord and the atlantoaxial ligament.

All the patients in our series were followed after surgery for more than 2 years. The follow-up period ranged from 28 to 85 months (mean follow-up, 39.5 [+ or -] 0.1 months). A cervical orthotic device (Miami-J or Philadelphia collar) was used in all of our patients for 4 to 6 weeks after surgery. An early (within 24 hours) postoperative CT scan was obtained to document the screw position. The follow-up examination was performed by a neurosurgeon not involved in the patients' initial management, whereas the late postoperative imaging studies were reviewed by an experienced neuroradiologist not involved in the patients' treatment. Adequate fusion was defined by using the following radiographic criteria: presence of bridging trabecular bone along the surface of the graft, C1-C2 joint without lucency or resorption of the graft, or C1-C2 motion on the dynamic radiographic views. Flexion and extension radiographs were obtained at 6 and 12 months after surgery. Complications associated with the malposition of the inserted screws or mechanical hardware failure were investigated by review of surgical and medical records.

Surgical Technique

The proposed technique by Magerl and Seemann (6) was used with a few modifications. The patient was initially positioned supine on a Jackson spinal operating table (OSI Inc, Union City, CA). Baseline intraoperative somato-sensory evoked potentials (SSEP) were obtained before induction. Fiberoptic, nasotracheal intubation was performed with the patient in the supine position. After induction, radiolucent sponges were used to keep the patient's mouth open. The operating table was rotated 180 degrees, and another set of SSEPs was obtained after positioning the patient. At this point, the importance of appropriate positioning must be emphasized. The patient's neck is in neutral position while his head is adequately flexed in a "military tuck" position; this will allow some manipulation during the procedure if this is necessary for better alignment of the C1-C2 unit. The procedure is routinely performed with the use of continuous SSEP monitoring. In cases of previous surgical intervention, and in which there was a halo vest, this was removed after positioning the patient supine while a dedicated physician maintained appropriate cervical alignment. Biplanar intraoperative fluoroscopic imaging was set for concomitant anteroposterior and lateral viewing during the screw insertion. The neck, from the occiput to the spinous process of C7 and the right posterior iliac crest, was prepared and draped in a standard sterile fashion to facilitate simultaneous graft harvesting and cervical dissection.

A midline longitudinal skin incision was made from the occiput (2 cm inferior to inion) to C7. The musculature was divided in the midline down to the tips of the spinous processes from C2 down to C7. With a standard subperiosteal technique, dorsal aspects of C1 through C7 vertebrae were exposed. The occiput and posterior arch of C1 were exposed as far laterally as possible without exposing the vertebral artery. The spinous processes, the laminae, and articular processes of C2 and C3 were also exposed subperiosteally; special care was taken not to violate the C2-C3 joint capsules (Fig. 1).

With the use of a small dissector, the cranial surface of the lamina and isthmus of C2 were carefully exposed bilaterally to the posterior capsule of the atlantoaxial joint. During this exposure, occasionally the venous plexus surrounding the greater occipital nerve was damaged, with consequent moderate venous bleeding, which was controlled by applying a hemostatic agent and by using bipolar coagulation. Special care must be taken not to place any pressure on the underlying dura and spinal cord during the maneuvers for controlling this bleeding. The greater occipital nerve and the surrounding vessels were gently retracted cranially with a curved dissector. At this point, the cartilage was removed from the joint with small curets. The medial and lateral aspects of the isthmus were exposed before screw-hole preparation. We have found that this maneuver minimizes the potential for drilling lateral into the vertebral artery or medial into the spinal canal. Landmarks used for defining the screw entry point were 2 mm above the center of the C2-C3 facet joint and 2 mm medial to the middle of the joint. This keeps the screws as far away from the vertebral artery as possible and at the same time allows the preparation of a second screw-hole, starting further laterally and angled more medially, if the initial screw trajectory is unsatisfactory (Figs. 2 and 3).

A high-speed drill was used to decorticate the entry site. With biplanar, real-time fluoroscopic imaging and direct visualization of the dorsal isthmus of C2, a 2.7-mm drill bit, powered by a high-speed drill (Anspach Companies, Palm Beach Gardens, FL) was used to initiate the screw-hole preparation. The drill was aimed at the anterior tubercle of C1, and was stopped at the posterior cortical surface of the anterior arch of C1. We occasionally used a Kocher clamp, placed on the spinous process of C2, and used to gently elevate the posterior elements of C2, facilitating the drilling at a greater angle in regard to the sagittal plane. A guiding K-wire was inserted in a trajectory of 10 to 15 degrees medially, and a cannulated screw was then inserted under fluoroscopic guidance (Figs. 4 and 5).

[FIGURE 1 OMITTED]

At the completion of the drilling, a short cortical screw was placed on top of the C2 isthmus to help determine the diameter of the permanent screw. The limiting factor is the transverse width of the isthmus directly below the C1-C2 joint. Additional information about the isthmus surface was obtained through preoperative CT. A 3.5-mm tap was used to continue the preparation of the hole before placement of the 3.5-mm diameter cortical screws, which are not self-tapping. A depth gauge was then inserted into the screw-hole to determine permanent screw length. The average screw length in this series was 40 mm (range, 34 to 44 mm) (Figs. 2 and 3). All of our patients underwent screw fixation with a permanent cortical screw measuring 3.5 mm in major diameter. Permanent cannulated screws were placed bilaterally in 21 cases (Fig. 6). In the remaining two patients, we were unable to insert a second screw because of technical difficulties. Bone was then used to part the C1 and C2 facet joints. The posterior arch of C1 and the C2 lamina and spinous process were decorticated with a high-speed drill, and the remaining matchstick bone graft was placed. The surgical wound was closed in anatomic layers. A drain insertion was necessary in only one case.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

Results

In our series, the mean operative time was 179.8 [+ or -] 0.7 minutes (range, 160 to 224 minutes). The mean blood loss during the procedure was 225.6 mL (range, 75 to 1,000 mL). In none of our cases were SSEP changes detected at the conclusion of the operation. The mean length of hospitalization was 3.4 [+ or -] 0.1 days (range, 2 to 11 days).

Regarding the occurrence of intraoperative or early postoperative complications, no spinal cord injury or nerve root injury occurred. No episode of intraoperative massive bleeding with suspected vertebral artery injury occurred. No patient had postoperative symptoms indicating vertebral artery injury. There were no traumatic aneurysms, arteriovenous fistulae, or vascular dissections.

Moreover, other complications not directly related to this particular surgical approach were encountered. Among them, superficial wound infection was the most common and the most bothersome in our series; two of our patients (8.7%) had such an infection, which was managed with intravenous antibiotics and surgical wound revision without the necessity of removing the implanted hardware. One patient (4.3%) had deep venous thrombosis, which was successfully managed by inserting a Greenfield filter in the inferior vena cava. Another elderly patient had postoperative development of left lower lobe atelectasis; this was adequately treated with antibiotics, with no further consequences.

Based on the postoperative CT scan, the position of the implanted screws was found to be satisfactory in 43 cases (21 patients), whereas in two cases (two patients), erroneous position of the screw was documented; in both of these cases, the implanted screws were positioned too medially, with subsequent extradural compression of the thecal sac at this level. Interestingly, both patients were neurologically intact and asymptomatic. On the basis of the patients' clinical neurologic examination, we decided not to reposition these suboptimally positioned screws. We found no screws penetrating the occipitoatlantal joint. It needs to be mentioned that we did not encounter any difficulties in defining the anatomic landmarks, nor in inserting the screws in the six cases with previous surgery.

In regard to the radiographic evidence of fusion in our series, the final follow-up assessment, consisting of flexion/extension radiography or occasionally polytomograms, showed solid osseous fusion (as defined earlier in our communication) in 19 patients (82.6%). In the remaining four patients (17.4%), pseudoarthrosis was found in their radiographic studies; in none of these patients, however, was abnormal motion detected in the obtained flexion/extension radiographs. In addition, failure of the implanted hardware (screw breakdown) was documented in two patients (8.7%). In both patients, a solid bony fusion was documented on the late postoperative radiographic evaluation, and their dynamic radiographs showed no abnormal motion at this level.

Finally, the preoperative symptomatology of our patients was resolved in the vast majority of them during the early postoperative period. Among the 18 patients with severe preoperative occipital pain (rated 8 to 9 out of a possible 10), only 1 (5.6%) complained of moderate pain (rated 5) of the same distribution during the first 4 postoperative weeks; his pain was eventually resolved with symptomatic conservative treatment. Three of our patients complained of some numbness in the occipital and high cervical areas, which resolved gradually over a 4-week period. All of our patients with preoperative myelopathic symptoms became symptom-free after surgery.

Discussion

Posterior C1-C2 transarticular screw fixation has been used successfully in the surgical management of atlantoaxial instability. (6-10,12-16) The greatest advantages of this technique are the excellent rotatory stability and the higher fusion rate compared with other posterior surgical stabilizing techniques. (9-11,15) In various studies, fusion rates are reported to range between 95 to 100%. (7-10,14) In our study, the fusion rate was 82.6%, lower than in most recently published clinical studies. It is important when attempting a comparison between fusion rates to clarify how fusion is defined, however. We strongly believe that when more strict radiographic criteria are followed for defining a solid fusion, then the accomplished fusion rate is generally lower than that reported in most of the published clinical series.

Several complications related to screw insertion have been reported, preventing surgeons from embracing this technique. These include screw malposition, pseudoarthrosis, dural tear, neurologic deficit, implant biomechanical failure, and compromise of the adjacent vascular structures. Among these, the most serious one is vertebral artery injury. Several studies have reported different vertebral artery injury rates, from 0 to 11.1%. (17) Fortunately, in our study, there was no incidence of vertebral artery damage. Injuries of the vertebral artery and spinal cord are considered to be highly correlated with screw malpositioning. Although the incidence of neurologic deficit from intraoperative vertebral artery injury is as low as 3.7%, efforts should be made to lower these risks. In a large multi-institutional retrospective study performed by Wright et al, (18) the incidence of documented vertebral artery injury was 2.4% and the incidence of a suspected injury was 1.7%. Based on a national retrospective survey, the risk of a vertebral artery injury was calculated to be 2.2% per inserted screw, whereas the risk of neurologic deficit secondary to an associated vertebral artery injury was 0.1% per inserted screw.

A number of strategies can minimize the risk of malpositioning the screw and consequently minimize the chance of any procedure-related complications. Dickman and Sonntag (7) described five key points that summarize the crucial steps of this procedure, and can help the performing surgeon to avoid an erroneous screw positioning. First, the screw insertion should be avoided if the course of the vertebral artery is tortuous, ectatic, or anomalous. This fact emphasizes the importance of obtaining adequate preoperative imaging of the C1-C2 area in great detail. Second, the insertion of transarticular screws should be avoided if the C1 lateral mass or C2 pars articularis is fractured or destroyed. Again, the importance of obtaining a high-resolution preoperative CT scan cannot be overemphasized. Third, the use of frameless stereotactic navigation can help the surgeon to judge the trajectory of the screw and avoid an erroneous screw placement. (7,19) The use of a surgical navigator adds to the safety of the procedure, regarding the passage of the screw through the isthmus of C2, which is by far the most narrow point in the screw trajectory. (20) It is apparent, though, that the surgeon must be aware of the lack of intersegmental registration in most of the commercially available frameless stereotactic systems. The exact knowledge of the regional anatomy cannot be substituted by the use of such a system. Fourth, intraoperative fluoroscopy provides real-time feedback to assess the trajectory, depth, and position of the screw. We have found that the use of biplanar fluoroscopy during the procedure is very time-efficient, even though in our first few cases we struggled with getting both C-arm units in place. However, in our hands, biplanar fluoroscopic guidance significantly diminishes the operative time and adds to the safety of the procedure. Fifth, the exposure of the C2 pars articularis and dorsal C1-C2 facet during surgery enables the surgeon to guide the trajectory of the screw under direct visualization. This operative maneuver can potentially lead to significant venous bleeding, as it did in a few of our cases, and such bleeding can be troublesome. However, it helped us to directly visualize the inserted screw trajectory, and thereby avoid screw malpositioning.

Another technical detail of great importance is related to the incomplete reduction of the atlantoaxial joint before the drill passage. We encountered this difficulty in our series in a couple of cases, but were able to reduce the atlantoaxial joint by applying some gentle traction on the spinous process of C2 by using a Kocher clamp. It is obvious that appropriate positioning of the patient in the beginning of the procedure is of paramount importance.

The issue of obtaining an early postoperative C-spine CT scan has remained controversial. (13,15) There are several groups that advocate against such an imaging study because they consider it unnecessary. (13) We strongly believe that an erroneously positioned transarticular screw, particularly in a patient with vague symptomatology, needs to be repositioned, or, if this is not technically possible, to be removed as soon as possible. We consider that an early postoperative CT scan with three dimensional reconstruction could provide the surgeon with valuable information regarding the position, and particularly the malposition, of the inserted screws, before the patient is mobilized. In the current clinical series, we faced the puzzling dilemma of having documented on the early postoperative CT scan suboptimally positioned screws in two asymptomatic patients. We decided not to reposition them based on their neurologic examination, but we closely followed both of these patients. The postoperative CT scan is reassuring for the surgeon, but in our clinical series did not alter the treatment plan in any of these patients. The issue will remain controversial, and the policies of several different institutions might in the near future dictate the necessity of this study.

Finally, the fact that in our series there were two cases (in two patients) of mechanical hardware failure is worrisome, and requires further investigation by the manufacturing companies. Florensa et al (14) also reported hardware failure in their large clinical series, and the authors are aware of several anecdotal instances of similar mechanical failure of the implanted hardware. This finding becomes even more important when this hardware is inserted in pediatric patients, in whom the mechanical stress of the implanted transarticular screws will be prolonged. The necessity of further biomechanical studies is apparent, so that the tensile strength of commercially available screws can be improved. In addition, a multi-institutional registry for reporting any mechanical failure of transarticular screws should be created in order to facilitate a better understanding of the long-term biomechanical properties of this hardware, which in turn would eventually lead to a better profile product.

Conclusion

C1-C2 posterior transarticular screw fixation is a biomechanically superior fixation technique that provides immediate rigid stability without the use of rigid external orthosis. This technique is particularly useful in cases of complex atlantoaxial instability. Meticulous preoperative planning and precise operative techniques are required for successful transarticular screw fixation, because the tolerance for error in this region of the spine is very limited. Ultimately, as emphasized in the relevant literature, it is up to the surgeon to weigh the advantages of this technique against the possible risk of screw malposition and neural or vascular injury, and to choose the technique most appropriate for each patient.

I have often regretted my speech, never my silence.
--Xenocrates

Acknowledgments

The authors thank Ms. Sadie B. Fennell for valuable assistance on the preparation of the manuscript and Mrs. Marjorie Lawrence for illustrations.

Accepted March 11, 2004.

Please see John Glaser's editorial on page 1030 of this issue.

References

1. Fielding JW. Cineroentgenography of the normal cervical spine. J Bone Joint Surg Am 1957;39:1280.

2. Dvorak J, Hayek J, Zehnder R. CT-functional diagnostics of the rotatory instability of the upper cervical spine: an evaluation on healthy adults and patients with suspected instability. Spine 1987;12:726.

3. Gallie WE. Fractures and dislocation of the cervical spine. Am J Surg 1939;46:495.

4. Brooks AL, Jenkins EB. Atlanto-axial arthrodesis by the wedge compression method. J Bone Joint Surg Am 1978;60:279-284.

5. Roosen K, Trauschel A, Grote W. Posterior atlanto-axial fusion: a new compression clamp for laminar osteosynthesis. Arch Orthop Trauma Surg 1982;100:27-31.

6. Magerl F, Seemann PS. Stable posterior fusion of the atlas and axis by transarticular screw fixation, in Kehr P, Weidner A (ed): Cervical Spine I. Vienna, Springer-Verlag, 1987, pp 322-327.

7. Dickman CA, Sonntag VK. Posterior C1-C2 transarticular screw fixation for atlantoaxial arthrodesis. Neurosurgery 1998;43:275-280.

8. Grob D, Jeanneret B, Aebi M, et al. Atlanto-axial fusion with transarticular screw fixation. J Bone Joint Surg (Br) 1991;73:972-976.

9. Haid RW, Subach BR, McLaughlin MR, et al. C1-C2 transarticular screw fixation for atlantoaxial instability: a 6-year experience. Neurosurgery 2001;49:65-70.

10. Mummaneni PV, Haid RW, Fiore AJ, et al. Posterior fixation options for the C1-C2 complex: wires, clamps, and screws. Contemp Neurosurg 2003;25:1-8.

11. Qu D, Jin D, Zhao W, et al. Biomechanical evaluation of atlantoaxial transarticular screw fixation technique. Chin J Traumatol 2000;15:89-92.

12. Reilly TM, Sasso RC, Hall PV. Atlantoaxial stabilization: clinical comparison of posterior cervi wiring technique with transarticular screw fixation. J Spinal Disord Tech 2003;16:248-253.

13. Haid RW. C1-C2 transarticular screw fixation: technical aspects. Neurosurgery 2001;49:71-74.

14. Florensa R, Noboa R, Munoz J, et al. Results of C1-C2 transarticular screw fixation in a series of 20 patients. Neurocirugia 2002;13:429-435.

15. Fountas KN, Smisson HF, Robinson JS Jr. C1-C2 transarticular screw fixation for atlantoaxial instability: a 6-year experience. Neurosurgery 2002;50:672-673.

16. Levy ML, McComb JG. C1-C2 fusion in children with atlantoaxial instability and spinal cord compression: technical note. Neurosurgery 1996;38:211-215.

17. Coric D, Branch CL Jr, Wilson JA, et al. Arteriovenous fistula as a complication of C1-2 transarticular screw fixation: case report and review of the literature. J Neurosurg 1996;85:340-343.

18. Wright NM, Lauryssen C. Vertebral artery injury in C1-C2 transarticular screw fixation: results of a survey of the AANS/CNS Section on Disorders of the Spine and Peripheral Nerves. J Neurosurg 1998;88:634-640.

19. Bloch O, Holly LT, Park J, et al. Effect of frameless stereotaxy on the accuracy of C1-2 transarticular screw placement. J Neurosurg 2001;95:74-79.

20. Neo M, Matsushita M, Yasuda T, et al. Use of an aiming device in posterior atlantoaxial transarticular screw fixation: technical note. J Neurosurg 2002;97:123-127.

RELATED ARTICLE: Key Points

* Transarticular C1-C2 screw fixation is favored as a biomechanically superior stabilization procedure.

* A retrospective series of 23 patients treated with this modality is presented.

* Postoperative complications and long-term results are analyzed.

* Indications for this technique are also discussed by the authors.

K. N. Fountas, MD, PHD, E. Z. Kapsalaki, MD, PHD, I. Karampelas, MD, V. G. Dimopoulos, MD, C. H. Feltes, MD, M. A. Kassam, MD, A. N. Boev, MS. III, K. W. Johnston, MD, FACS, H. F. Smisson, MD, FACS, E. C. Troup, MD, and J. S. Robinson Jr, MD, FACS

From the Departments of Neurosurgery and Radiology, The Medical Center of Central Georgia, Mercer University School of Medicine, Macon, GA.

The authors have no commercial or proprietary interest in any drug, device, or equipment mentioned in this article.

Reprint requests to Dr. Kostas N. Fountas, 840 Pine St, Suite 880, Macon, GA 31201. Email: [e-mail]