OBJECTIVEThe purpose of this study was to compare stability of injectable hollow pedicle screws with different numbers of holes using different volumes of polymethylmethacrylate (PMMA) in severely osteoporotic lumbar vertebrae and analyze the relationship between screw stability and distribution and volume of PMMA.METHODSForty-eight severely osteoporotic cadaveric lumbar vertebrae were randomly divided into 3 groups-groups A, B, and C (16 vertebrae per group). The screws used in group A had 4 holes (2 pairs of holes, with the second hole of each pair placed 180° further along the thread than the first). The screws used in group B had 6 holes (3 pairs of holes, placed with the same 180° difference in position). Unmodified conventional screws were used in group C. Each group was randomly divided into subgroups 0, 1, 2, and 3, with different volumes of PMMA used in each subgroup. Type A and B pedicle screws were directly inserted into the vertebrae in groups A and B, respectively, and then different volumes of PMMA were injected through the screws into the vertebrae in subgroups 0, 1, 2, and 3. The pilot hole was filled with different volumes of PMMA followed by insertion of screws in groups C0, C1, C2, and C3. Distributions of PMMA were evaluated radiographically, and axial pull-out tests were performed to measure the maximum axial pullout strength (Fmax).RESULTSRadiographic examination revealed that PMMA surrounded the anterior third of the screws in the vertebral bodies (VBs) in groups A1, A2, and A3; the middle third of screws in the junction area of the vertebral body (VB) and pedicle in groups B1, B2, and B3; and the full length of screws evenly in both VB and pedicle in groups C1, C2, and C3. In addition, in groups A3 and B3, PMMA from each of the screws (left and right) was in contact with PMMA from the other screw and the PMMA was closer to the posterior wall and pedicle than in groups A1, A2, B1, and B2. One instance of PMMA leakage was found (in group B3). Two-way analysis of variance revealed that 2 factors-distribution and volume of PMMA-significantly influenced Fmax (p < 0.05) but that they were not significantly correlated (p = 0.078). The Fmax values in groups in which screws were augmented with PMMA were significantly better than those in groups in which no PMMA was used (p < 0.05).CONCLUSIONSPMMA can significantly improve stability of different injectable pedicle screws in severely osteoporotic lumbar vertebrae, and screw stability is significantly correlated with distribution and volume of PMMA. The closer the PMMA is to the pedicle and the greater the quantity of injected PMMA used, the greater the pedicle screw stability is. Injection of 3.0 mL PMMA through screws with 4 holes (2 pair of holes, with the screws in each pair placed on opposite sides of the screw) produces optimal stability in severely osteoporotic lumbar vertebrae.