Atlantoaxial transarticular fixation, also called the Magerl technique, is said to be the most robust biomechanical method of fixation of the atlantoaxial vertebrae. However, the procedure carries a risk of spinal cord and vertebral artery injury during the insertion process, especially in patients with a high-riding vertebral artery. In this study, a computed tomography (CT)-based navigation system was used for preoperative planning and insertion. This investigation sought to determine the rate and direction of screw perforation as well as the incidence of screw loosening in computer-assisted atlantoaxial transarticular fixation. Sixty patients (31 men and 29 women; mean ± SD age: 65.3 ± 19.6 years) who received atlantoaxial transarticular screw insertion with preoperative CT navigation were analyzed. We investigated screw position and loosening by CT at the final follow-up. Of the 108 screws inserted, the rate of Grade 2 or higher perforation was 4.6% (5/108). Nine of 81 (11.1%) screws inserted into the 44 patients who were followed for at least 6 months showed loosening. Logistic regression analysis revealed that unilateral insertion (odds ratio: 8.50, 95% confidence interval: 1.53-47.2, P = 0.014) was significantly associated with the incidence of screw loosening. The screw perforation rate of Grade 2 or higher in computer-assisted atlantoaxial transarticular screw fixation was 4.6%, with comparable frequencies of perforation direction. Unilateral insertion was a significant independent factor associated with screw loosening, which occurred in 11.1% of insertions. Spine surgeons should follow up with patients with caution because screws with unilateral insertion are prone to loosening.