Background: Posterior fixation practices after endonasal odontoidectomy have moved away from
occipitocervical fixations (OCF) to atlantoaxial fusions (AAF), and endonasal decompressions
have transitioned from complete anterior C1 arch resections, to partial resections,
preserving atlas' structural integrity. Furthermore, complete resection of the retropharyngeal
soft tissues has evolved to vascularized rhinopharyngeal (RP) flap reconstructions.
Objective: We sought to evaluate the impact this transition in practices had on clinical outcomes,
based on our institutional experience.
Methods: Retrospective review of endoscopic endonasal odontoidectomy cases with instrumented
fusion between 2008 and 2020. Electronic medical records were queried for demographic,
operative, and follow-up data. Pre- and postoperative CT and MR images were evaluated
and ventral triangular area of decompression calculated ([Fig. 1 ]).
Results: Thirty-four patients underwent endonasal odontoidectomy and fusion for basilar invagination
(n = 12) or rheumatoid pannus (n = 22); 19 received OCF and 15 AAF, with average follow-up of 34 months. Myelopathy
was the most common presenting feature, 23/34 (68%), followed by dysphagia, 14/34
(41%). A rhinopharyngeal flap was used in 12/34 (35%). Of RP patients, 3/12 had preexisting
dysphagia none of which worsened postoperatively. Rates of infection were similar
with or without RP. Rates of CSF leak, wound complications, and chronic neck pain
were similar between OCF and AAF. There was no increased risk of hardware failure,
deformity, adjacent segment disease (ASD), or redo cervical operation with AAF compared
with OCF. There was an increased risk of postop dysphagia in patients with OCF (8/19;
2/8 had pre-op dysphagia) compared with AAF (1/15; the one had pre-op dysphagia; 42
vs. 7%; p = 0.047). Risk of tracheostomy was 26% among the OCF cohort versus 7% in AAF patients
(p = 0.13). Twenty-four patients had complete C1 arch resection while 10 were incomplete.
There was no difference in degree of ventral decompression of the cervicomedullary
junction with complete versus incomplete arch of C1 resection. The average ventral
triangular area of compression pre- and postoperatively among patients with complete
arch removal was 2.65 and 1.24 cm2 , respectively, whereas in patients with incomplete removal the values were 3.00 and
1.49 cm2 , respectively. Within our follow-up period was no significant difference between
rates of hardware failure, deformity, ASD, or redo cervical operation between OCF
and AAF with or without complete resection of the anterior arch of C1.
Conclusion: AAF does not confer increased risk of hardware failure, deformity, ASD, or revision
compared with OCF, while the latter is associated with increased risk for dysphagia.
AAF should therefore be performed preferentially over OCF when feasible. Partial resections
of the anterior arch of C1 did not lessen degree of decompression and may help maintain
atlas structural integrity. These changes combined with use of an RP flap represent
an evolution of the overall management of irreducible odontoid disease.
Fig. 1 Ventral triangular compression calculation.
