Pedicle screws are commonly used for posterior stabilization of the spine. When used in deformed or degenerated segments, the pedicle screws are often not fully inserted into the bone, but instead the threaded portion is exposed by 1 or 2 threads to accommodate rod placement and ensure alignment between the tulip of the screw and the rod. However, broken pedicle screws have been reported with the use of this method. The aim of this study was to determine how the fatigue life of the screw is affected by not fully inserting the screw into the bone. Spinal constructs were evaluated in accordance with ASTM F1717. The following three screw positions were subjected to compression bending fatigue loading; (i) pedicle screw fully inserted in the test block with no threads exposed (EXP-T0), (ii) pedicle screw inserted with one thread exposed outside the test block (EXP-T1), (iii) pedicle screw inserted with two threads exposed outside the test block (EXP-T2). Corresponding finite element models FEM-T0, FEM-T1 and FEM-T2 were also constructed and subjected to the same axial loading as the experimental groups to analyze the stress distribution in the pedicle screws and rods. The results showed that under a 190 N axial load, the EXP-T0 group survived the full 5 million cycles, the EXP-T1 group failed at 3.7 million cycles on average and the EXP-T2 groups failed at 1.0 million cycles on average, while the fatigue strength of both the EXP-T1 and EXP-T2 groups was 170 N. The constructs failed through fracture of the pedicle screw. In comparison to the FEM-T0 model, the maximum von Mises stress on the pedicle screw in the FEM-T1 and FEM-T2 models increased by 39% and 58%, respectively. In conclusion, this study demonstrated a drastic decrease in the fatigue life of pedicle screws when they are not full inserted into the plastic block.