METHODS A retrospective radiographic and medical record analysis of 58 patients. OBJECTIVE To describe the incidence and consequences of cervical spinal deformity and instability after multilevel laminectomy in adult patients with myelopathy caused by cervical spondylosis and to determine the usefulness of preoperative dynamic roentgenographic films in the prevention of postoperative destabilization. BACKGROUND Extensive cervical laminectomy has been widely used in the treatment of progressive myelopathy secondary to stenotic conditions. Complications of this procedure, including spinal instability, accelerated spondylotic changes, postoperative spinal deformity, and constriction of the dura mater by formation of extradural scar tissue formation have been recognized. However, the frequency of these complications is probably overestimated, and their effect on clinical outcome remains unknown. METHODS Fifty-eight patients older than 30 years who underwent a laminectomy at more than three levels without fusion for myelopathy secondary to cervical spondylosis were reviewed retrospectively with an average follow-up of 3.6 years. Functional results were evaluated according to the Japanese Orthopaedic Association's scoring system. Lateral views in neutral position, in flexion, and in extension of the preoperative cervical roentgenograms were analyzed in comparison with the last follow-up films to identify the changes in the curvature of the cervical column, in the range of motion of the neck, and in the intervertebral angular mobility and anteroposterior displacement of the vertebral bodies and finally to quantify the incidence of spinal instability. RESULTS In 18 patients (31%), postoperative changes in the type of cervical spine curvature developed. Fifteen patients (25%) had destabilization at one or more levels. Deformities of the cervical spine occurring after surgery do not appear to cause symptoms or neurologic abnormalities. Destabilization required repeat surgery in 3 patients. All the levels appearing to be destabilized on the postoperative films were hypermobile on the preoperative dynamic radiographs. Preoperative olisthesis Without hypermobility is not a factor of risk in postoperative destabilization. CONCLUSIONS The use of preoperative dynamic radiographs should improve the selection of patients undergoing laminectomy for the treatment of multilevel cervical cord compression. Dynamic radiographs may also reinforce the need for such adjunctive procedures as fusion and instrumentation, to prevent postoperative destabilization. Preoperative olisthesis with hypermobility in sagittal or horizontal planes must be fused and instrumented.