An animal model of anterior and posterior column Instability was developed to allow in vivo observation of bone remodeling and arthrodesis after spinal instrumentation. After an initial anterior and posterior destabilizing lesion was created at the L5-L6 vertebral levels in 42 adult beagles, various spinal reconstructive surgical procedures were performed—with or without bilateral posterolateral bone grafting, and with or without spinal instrumentation (Harrington distraction; Luque rectangular, or Cotrel-Dubousset transpedlcular methods). After 6 months' postoperative observation, there was a significantly improved probability of achieving a spinal fusion if spinal instrumentation had been used (P=0.058). Nondestructive mechanical testing after removal of all metal Instrumentation in torsion, axial compression, and flexion revealed that the fusions performed in conjunction with spinal instrumentation were more rigid (P<0.05). Quantitative hlstomorphometry showed that the volumetric density of bone was significantly lower (le, device-related osteoporosis occurred) for fused versus unfused spines; and Harrington- and Cotrel-Dubousset-instrumented dogs became more osteoporotic than the other three groups. The rigidity of spinal instrumentation led to device-related osteoporosis (stress shielding) of the vertebra. However, as the rigidity of spinal Instrumentation increased, there was an increased probability of achieving a successful spinal fusion. The Improved mechanical properties of spinal instrumentation oh spinal arthrodesis more than compensate for the occurrence of device-related osteoporosis In the spine.