Background ContextA high-riding vertebral artery (HRVA) can deviate too medially, too posteriorly, or too superiorly to allow the safe insertion of screws. However, it is unknown whether the presence of a HRVA is associated with morphological changes of the atlantoaxial joint. PurposeTo investigate the association between HRVA and atlantoaxial joint morphology in patients with and without HRVA. Study DesignA retrospective case–control study and finite element (FE) analysis. Patient SampleA total of 396 patients with cervical spondylosis underwent multi-slice spiral computed tomography (MSCT) of cervical spine at our institutions from 2020 to 2022. Outcome MeasuresA series of atlantoaxial joint morphological parameters, including C2 lateral mass settlement (C2 LMS), C1–2 sagittal joint inclination (C1–2 SI), C1–2 coronal joint inclination (C1–2 CI), atlanto-dental interval (ADI), lateral atlanto-dental interval (LADI), and C1–2 relative rotation angle (C1–2 RRA) were measured, and lateral atlantoaxial joints osteoarthritis (LAJs-OA) was recorded. The stress distribution on the C2 facet surface under different torques of flexion-extension, lateral bending, and axial rotation was analyzed by FE models. A 2-Nm moment was applied to all models to determine the range of motion (ROM). MethodsA total of 132 consecutive cervical spondylosis patients with unilateral HRVA were enrolled in the HRVA group, and 264 patients without HRVA matched for age and sex were enrolled in the normal (NL) group. Atlantoaxial joint morphological parameters were compared between two sides of C2 lateral mass within HRVA or NL group, and between HRVA and NL groups. A 48-year-old woman with cervical spondylosis without HRVA was selected for cervical MSCT. A three-dimensional (3D) FE intact model of the normal upper cervical spine (C0-C2) was created. We established the HRVA model by simulating atlantoaxial morphological changes of unilateral HRVA with FE method. ResultsThe C2 LMS was significantly smaller on the HRVA side than that on the non-HRVA side in the HRVA group, but C1–2 SI, C1–2 CI, and LADI on HRVA side were significantly larger than those on non-HRVA side. There was no significant difference between left and right sides in the NL group. The difference in C2 LMS (d-C2 LMS) between HRVA side and non-HRVA side in the HRVA group was larger than that in the NL group (P < 0.05). Meanwhile, the differences in C1–2 SI (d-C1/2 SI), C1–2 CI (d-C1/2 CI), and LADI (d-LADI) in the HRVA group were significantly larger than those in the NL group. The C1–2 RRA in the HRVA group was significantly larger than that in the NL group. Pearson correlations showed that d-C1/2 SI, d-C1/2 CI, and d-LADI were positively associated with d-C2 LMS (r=0.428, 0.649, 0.498, respectively, p<.05 for all). The incidence of LAJs-OA in the HRVA group (27.3%) was significantly larger than that in the NL group (11.7%). Compared with the normal model, the ROM of C1-2 segment declined in all postures of the HRVA FE model. We found a larger distribution of stress on the C2 lateral mass surface of the HRVA side under different moment conditions. ConclusionsWe suggest that HRVA affects the integrity of the C2 lateral mass. This change in patients with unilateral HRVA is associated with the nonuniform settlement of the lateral mass and an increase in the lateral mass inclination, which may further affect the degeneration of the atlantoaxial joint because of the stress concentration on the C2 lateral mass surface.