Purpose of the study: The Harms technique of stabilizing C1-C2 by fixation with polyaxial screws and rods is a further option for atlantoaxial fixation from the dorsal approach. Harms and Melcher published this method in 2001, but the operation had first been performed by Harms in August 1997. The aim of this study is to evaluate the first results and try to assign the Harms C1-C2 fixation an appropriate standing in the in broad range of options for stabilization of the atlantoaxial complex.
Material: Between December 2002 and January 2004 we carried out the Harms fixation of C1-C2 on 22 patients admitted to the Department of Spine Surgery, Motol University Hospital, 2nd Medical Faculty in Prague. Out of these, 18 patients were included in this study, 10 men and 8 women between 23 and 84 years of age (average, 55.4 years) followed-up longer than 6 months. In 14 patients we used the Harms technique as a permanent fixation of C1-C2 in order to achieve atlantoaxial arthrodesis and, in four patients, we applied it only for a period of 4 to 6 months without the use of bone grafts or their substitutions. We employed the permanent fixation to treat the following conditions: fracture of the atlas in three patients, type IIA comminuted fracture of the dens base in three patients, fracture of C2 categorized as "other" in two patients, atlantoaxial vertical instability in one patient with rheumatoid arthritis, malunion of the fractured dens in one patient, and complicated trauma to C1-C2 in four patients. The temporary fixation was used for type III displaced fractures of the dens in two and fixed atlantoaxial rotatory dislocations also in two cases. Only one patient showed signs of Frankel C neurological deficit on admission, the rest were without neurological findings.
Methods: All screws were inserted under an image intensifier always in lateral projection. First we retracted the greater occipital nerve in a caudal direction towards C2 with a fine raspatory and, using an awl, marked the entry point in the C1 lateral mass; a pilot hole, reaching through the anterior cortical bone, was made with a 2.5 mm drill. It followed a straight or slightly convergent trajectory in an anterior-posterior direction and parallel to the plane of the C1 posterior arch in the sagittal direction. Individual anatomical variations in the atlantoaxial complex of every patient were respected. The hole was tapped through the entire vertebral body, with the exception of osteoporous bone in which only the posterior cortical bone was treated with a screw tap. At this stage profuse bleeding usually arose from dissection around the epidural venous plexus along the C1-C2 joint. This was effectively controlled by a quick insertion of a screw and compression of the venous plexus with the screw head. To control bleeding by bipolar electrocautery is difficult and is always associated with a risk of nerve injury. Screws 3.5 mm thick, with polyaxial heads, were inserted bicortically into the lateral mass of C1. Subsequently, the intervertebral C2-C3 joint was localized and its medial border in the spinal canal was palpated. The entry point for placement of a C2 pedicle screw was marked with an awl at the point of intersection at a distance of 2 mm from the medial border and 5 mm from the caudal border of the C2 articular process. Under an X-ray intensifier in lateral projection, a hole was drilled approximately parallel to the screws inserted in C1, i. e., at an angle of 20 to 30 degrees cranially, up to and through the anterior cortical bone. In the transversal plane, the screws were situated in a convergent direction at an angle of 20 to 25 degrees. After all screws had been inserted, we reduced the antlantoaxial complex in the correct anatomical position by manipulating the patient's head or by directly adjusting the screws. Connecting 3.0-mm rods were then applied and fastened by cap nuts or inner nuts according to the instrumentation used.
Results: Operative time ranged from 35 to 155 min, with an average of 81 min. Intra-operative blood loss ranged from 50 to 1500 ml, with an average of 560 ml. The X-ray intensifier was used for a period of 0.4 to 2.6 min, with an average of 0.9 min. A total of 36 screws were inserted in the atlas; their length ranged from 16 to 34 mm (average, 30.6 mm). All screws were positioned correctly in the C1 lateral mass; two screws did not reach up to the anterior cortical bone and one protruded over it, but without causing clinical problems. Thirty-six screws were inserted in the axis. Their length ranged from 28 to 36 mm (average, 31.7) mm). Twenty-seven screws were correctly applied through the isthmus into the C2 anterior cortical bone, three were too short to reach it and five were placed too close to the vertebral artery canal. Of these, two protruded into the artery canal, but without clinical consequences. One screw inserted too medially passed into the spinal canal, but this also was without clinical response. Of the 36 screws inserted in C2, three (8.3 %) were malpositioned. Bony fusion at C1-C2 was the goal of this operation in 14 patients. At 6 weeks post-operatively, it was achieved in two patients, at 12 weeks in 12 patients and at 6 months in all 14 patients. The C1-C2 segment was stable at 12 weeks in all 18 treated patients. Four patients reported restriction of motion in rotation by 10 to 25 % after removal of the instrumentation.
Discussion: Operative time, longer at the beginning than with the Magerl technique, gradually shortened to between 45 and 60 min. Similar trends were seen when intra-operative blood loss and X-ray exposure were evaluated. Using the Harms and Melcher procedure we saved the greater suboccipital nerve. In contrast to these authors, however, we did not resect the atlantoaxial joint. Solid fusion was achieved in all our patients. Of the total of 72 screws inserted, only three (4.2 %) were assessed as malpositioned; however, when related to the 36 screws inserted in C2, this was 8.3 %, which indicates that insertion of screws in C2 was more difficult. We did not observe any clinical consequences in any of these cases.
Conclusions: The Harms fixation of C1-C2 is a very effective technique for stabilizing the atlantoaxial complex. It enables us to provide temporary fixation without damage to atlantoaxial joints and to reduce the vertebrae after the screws and rods had been inserted, which is unique. These advantages compensate for a higher cost of the implant.