During the last 6 years 24 patients with cervical or thoracic spinal cord injuries and a severe reflex neuropathic bladder underwent selective dorsal sacral rhizotomy with the aid of intraoperative neurostimulation and urodynamic monitoring. Preoperative and postoperative evaluation was available in 17 patients. Followup ranged from 2 months to 5 years (mean 32 months). Mean bladder capacity increased significantly after rhizotomy (from 148 +/- 28.1 to 377 +/- 52.9 ml., p less than 0.001), as did mean volume to first contraction (from 99 +/- 28.6 to 270 +/- 39.3 ml., p less than 0.001). No significant changes in bowel or erectile function were noted. Continence was improved in 94%, with 14 patients remaining completely dry and voiding with electrical stimulation or intermittent self-catheterization. The long-term results of selective sacral rhizotomy compare favorably to more aggressive alternatives, such as augmentation cystoplasty or urinary diversion.