ObjectivesIt is now understood that pedicle screw loosening at the implant‐bone interface can lead to poor screw–bone interface purchase and decreased fixation stability. Previous biomechanical tests used cadaveric vertebrae and pull‐out or torque loads to assess the effect of the insertional direction of pedicle screws on screw loosening. However, these tests faced challenges in matching biomechanical differences among specimens and simulating in vivo loads applied on pedicle screws. This study aimed to evaluate the effect of the insertional direction of pedicle screws on screw loosening using tension‐compression‐bending loads and synthetic bone vertebrae.MethodsPolyaxial pedicle screws were inserted into nine synthetic bone vertebrae in three directions (three samples per group): cranial, parallel, and caudad (−10°, 0°, +10° of the pedicle screw rod to the upper plane of the vertebra, respectively). Pedicle screws in the vertebrae were loaded using a polyethylene block connected to a material testing machine. Tension‐compression‐bending loads (100N–250N) with 30,000 cycles were applied to the pedicle screws, and displacements were recorded and then cycle‐displacement curve was drawn based on cycle number. Micro‐CT scans were performed on the vertebrae after removing the pedicle screws to obtain images of the screw hole, and the screw hole volume was measured using imaging analysis software. Direct comparison of displacements was conducted via cycle‐displacement curve. Screw hole volume was analyzed using analysis of variance. The correlation between the displacement, screw hole volume and the direction of pedicle screw was assessed by Spearman correlation analysis.ResultsThe smallest displacements were observed in the caudad group, followed by the parallel and cranial groups. The caudad group had the smallest screw hole volume (p < 0.001 and p = 0.009 compared to the cranial and parallel groups, respectively), while the volume in the parallel group was greater than that in the cranial group (p = 0.003). Correlation analysis revealed that the insertional direction of the pedicle screw was associated with the displacement (p = −0.949, p < 0.001) and screw hole volume (p = −0.944, p < 0.001).ConclusionStrong correlations were found between the insertional direction of the pedicle screw and relevant parameters, including displacement and screw hole volume. Pedicle screw insertion in the caudad direction resulted in the least pedicle screw loosening.