BACKGROUND CONTEXT: Researchers studying early pedicle screw designs have suggested that pullout strength and insertional torque are correlated. For surgeons using pedicle screws, insertional torque is widely believed to be a good predictor of pullout strength and initial stability of the screw and construct. How appropriate is this assumption when applied to new screw and thread designs? PURPOSE: This study investigated the correlation between insertional torque and pullout strength of three different pedicle screw designs, with different insertional torque characteristics. We hypothesized that a significant increase in insertional torque would indicate a commensurate increase in pullout strength. STUDY DESIGN: Biomechanical analysis of instrumented vertebral specimens. METHODS: Calf lumbar vertebra were prepared and instrumented with one of three pedicle screws. Pilot hole preparation was standardized and coaxial orientation was confirmed by direct inspection. Screws did not penetrate the pedicle cortex or abut or penetrate the anterior vertebral cortex. Any specimen with pedicle wall breach was discarded. The pedicles were instrumented with one of three screws: 1) 7.5 x 40 mm conical, asymmetric progressive thread (Xia; Stryker Spine, Allendale, NJ), 2) 7.5 x 40 mm conical with traditional V-shaped thread (Osteonics, Stryker Spine, Allendale, NJ)) or 3) 6.5 x 40 mm cylindrical with V thread (Osteonics, cylindrical). Paired testing allowed individual screws to be directly compared with a contralateral "control." Insertional torque and peak torque values were recorded for each rotation up to full insertion. Pullout testing was conducted at a rate of 1 mm/minute. Load-displacement data were recorded at 20 Hz. Stiffness was considered the slope of the most linear part of the curve before the yield point. RESULTS: Peak loads for 7.5 conical Xia screws measured 1,783+/-589.1 N compared with 1,943.0+/-625.8 N for 7.5 conical Osteonics screws and 1,641.0+/-356.7 N for 6.5 cylindrical Osteonics screws. The peak insertional torque values were 6.7+/-1.9 Nm (158% greater than control), 4.5+/-1.1 Nm (73% greater than control) and 2.6+/-0.7 Nm, respectively. Insertional torques for Xia screws were significantly greater than conical (p=.001) and cylindrical Osteonics screws (p<.0001), and insertional torques for Osteonics conical screws were significantly greater than those of cylindrical screws (p<.0001). Although pullout loads for the conical Osteonics screws were consistently higher than either the Xia or Osteonics cylindrical screws, the differences were not significant (p>.05). There was no significant correlation between pullout strength and insertional torque (p>.05). CONCLUSIONS: This unexpected result is best explained by the progressively narrowing flutes of the Xia screw, which compact the trabeculae into a smaller volume as the screw nears full insertion. The trapezoidal threads also increase contact with the cortical surface area and compress trabeculae toward the cortex, thus creating greater friction and higher torque values. This increase in torque did not translate into a commensurate increase in pullout strength, where trabeculae fail in shear.