Although an anterior approach with anterior lateral screw fixation has been developed for stabilizing the thoracolumbar spine clinically, screw loosening still occurs. In this novel in vitro study, we attempted to elucidate the optimal screw position in the lateral lumbar vertebra and the effect of bicortical fixation. A total of 72 fresh-frozen lumbar vertebrae from L1-6 were harvested from 12 mature pigs and randomly assigned to two modalities: bicortical fixation (n = 36) and unicortical fixation (n = 36). Six groups of screw positions in the lateral vertebral body in each modality were designated as central-anterior, central-middle, central-posterior, lower-anterior, lower-middle, and lower- posterior; 6 specimens were used in each group. The correlations between screw fixation modalities, screw positions and axial pullout strength were analyzed. An appropriate screw trajectory and insertional depth were confirmed using axial and sagittal X-ray imaging prior to pullout testing. In both bicortical and unicortical fixation modalities, the screw pullout force was significantly higher in the posterior or middle position than in the anterior position (p < 0.05), and there was no significant differences between the central and lower positions. The maximal pullout forces from the same screw positions in unicortical fixation modalities were all significantly lower, decreases that ranged from 32.7 to 74%, than those in bicortical fixation modalities. Our study using porcine vertebrae showed that screws in the middle or posterior position of the lateral vertebral body had a higher pullout performance than those in the anterior position. Posteriorly positioned lateral vertebral screws with unicortical fixation provided better stability than anteriorly positioned screws with bicortical fixation.