PubMed1993
Biomechanical evaluation of twelve different spinal devices in vitro employing pedicle screws was performed using static (n = 5) and cyclical testing (n = 3) parameters. In general, the rank order of implant failures was similar between static and cyclical tests, performed at 600 N compressive load, 5 Hz, and 1 million cycles. The mean number of cycles to failure was higher for spinal instrumentation systems employing longitudinal rods than those using plates (ANOVA F = 16.94, P < .001). At 600 N, the compact Cotrel-Dubousset, TSRH, and ISOLA rod systems demonstrated mean cycles to failure ranging from 200,000 to 800,000 cycles. The remaining devices including Dyna-lok, Kirschner plate, and VSP devices had failures ranging from 50,000 to 210,000 cycles. Polyethylene cylinders representing vertebral bodies were used to eliminate the problems of biologic deterioration encountered with cadaveric spines (a full cyclical test to 1 million cycles required 56 hours), and thus to provide analysis of the weak portion of each spinal system. The failure of eleven of the twelve spinal systems occurred by fracture of a pedicle screw, most commonly at the junction of the upper screw thread and the collar (Kirschner, AO fixator, standard CD, ISOLA, and TSRH). However, in Dynalok and VSP systems, fracture of the threaded portion of the screw just posterior to the integral nuts was the most common screw fracture location. The compact CD system was the only spinal implant that consistently failed by fracture of the longitudinal spinal member (rod). The fatigue life of rod based systems was longer than plate based systems. These studies confirm the importance of anterior column load sharing (vertebral body, corpectomy bone graft) as the mean bending strength demonstrated by these implant systems was not inordinately high using this "worst case scenario" model.