A New Screw Pullout Testing Under Effects of Stress Relaxation Serkan Inceoglu, Robert F. McLain, Suleyman Cayli, Cumhur Kilincer, Lisa Ferrara Spine Research Laboratory, The Cleveland Clinic Spine Institute, The Cleveland Clinic Foundation 1945 E97th Street, Cleveland, OH 44195 USA Introduction A compression or disruption of the spinal cord or nerve roots caused by a vertebral fracture can be extremely painful. Reconstructive surgery might then be needed to create a suitable environment in order for bony incorporation to take place at the fracture site and concomitant relief of compression on the nerves. Pedicle screw fixation is one of the most commonly used methods in spinal surgery. A pedicle fixation system is composed of at least four screws inserted into vertebral bodies through the pedicles, tubular cancellous bone surrounded by a cortical bone shell, and two connecting rods. Despite of its success, failure of pedicle screw fixation through loosening or pullout has been, however, reported. Bone is a viscoelastic material. The mechanical and viscoelastic properties of bone have been well investigated by many researchers; however, there is no information available on the viscoelastic behavior of the bone-screw interface. Understanding the mechanics of the bone-screw interface is clinically very significant for developing proper implant design criterions and choosing the right implant for surgery. The mechanical properties of the bone-screw interface are studied in-vitro by means of standard axial pullout test. The standard pullout test is the application of an axial load at a constant rate to the screw inserted into vertebra. This test gives information on the initial fixation strength and stiffness of the interface; however, it does not account for the effects of time-dependent properties of the interface on the biomechanical performance. This study was designed to observe the effects of stress relaxation at the bone-pedicle screw interface on its mechanical properties. Methods Specimen Preparation Three lumbar (T13-L6) spines were obtained from 14-week-old calves. The spines were cleaned of soft tissue, separated into individual levels of vertebra and kept frozen after wrapping in moist paper towels until testing. Hole preparation for screw insertion were done with respect common surgical procedures[1]. Pedicle screws in size of 6.5x40 mm (Xia, Stryker Spine, New Jersey) were inserted bilaterally into each vertebral specimen. The specimens instrumented were embedded into customized gripping fixtures using a liquid Cerrobend metal (Cerro Metal Products Co., Bellefonte, PA) to provide adequate gripping forces. Care was taken to keep the screws and screw insertion region free of metal embedding material. Biomechanical Testing An MTS Alliance RT/10 materials testing machine (MTS Corp., Eden Prairie, MN) was used for testing. The specimens were secured into the testing machine by means of custom fixtures. A universal joint attached to the superior actuator of the machine anchored the screw head through the use of an adapter. The inferior fixture was a platen with rigid clamps that was secured to a second universal joint. The fixtures utilized universal joints to allow the majority of axial force through the axis of the screw and eliminating any residual stresses due to rigid clamping. Each specimen was tested to failure, which was defined as the highest load that the bone-screw interface could bear. Pullout protocols, (i.e., the stress relaxation model and standard pullout model), were selectively randomized to the right and left pedicles of each vertebrae. In the standard pullout model defined by American Society for Testing and Materials Standard F543-01 A3 and commonly used pullout model in spinal biomechanics, served as a control model. In this situation, the screws were withdrawn continuously under a constant rate until the load exceeded the failure point [1]. In the stress relaxation model, the rate of withdrawal was the same as in control. However the withdrawal was paused for 1000s between each 0.5mm advancement of the actuator until the load exceeded the failure point. The rest time was chosen as 1000s not only due to practicality of the experiment but also since it was reported that significant amount of relaxation in bovine bone was completed in first 10 3 -1.5x10 3 s [2]. Testing was repeated for three different pullout rates: 1 mm/min, 5 mm/min and 25 mm/min. During testing load-displacement data were recorded at a data acquisition rate of 20Hz. Stiffness, “displacement-to-failure” and “energy-to-failure” data were calculated. The displacement-to-failure was calculated from the displacement of crosshead and the energy-to-failure was calculated as the area under the load-displacement curve. Thus, displacement-to-failure and energy-to-failure data were calculated only between 150N load and the failure point which was the maximum load reached during testing of each screw for an unbiased comparison. Descriptive statistics and two-way ANOVA for repeated measures with Bonferroni post-test detected differences between the groups at the 95% confidence level.