Rotational changes of the vertebral-pelvic axis following Cotrel-Dubousset instrumentation.

Ten consecutive patients with adolescent idiopathic scoliosis and King-Moe curve Types II and III, scheduled consecutively for Cotrel-Dubousset instrumentation, underwent pre- and postoperative computed tomography scans with axial slices through each vertebra, and including the pelvis. Vertebral rotation was measured and referenced to the pelvis. Average derotation of the thoracic apex after surgery was 9%. King-Moe Type II curves tended to derotate more successfully (average 26% improvement), while Type III curves derotated very little, if at all (average 1.3% worsening of the rotational deformity). Type II curves often showed segmental rotational changes outside the levels of instrumentation, while Type III curves did not; more frequently the spinal-pelvic axis rotated en bloc. It appears, therefore, that Cotrel-Dubousset instrumentation does not consistently or predictably derotate the thoracic apex relative to the pelvis, and coronal plane correction may only be apparent, due to transmitted torque and rotation of the entire spinal-pelvic axis.