We began this study to determine the efficacy and safety of intrathecal baclofen (ITB) delivered by a programmable pump for the treatment of severe spasticity of spinal cord origin. One hundred fifty two patients with severe spasticity of spinal origin, refractory to oral baclofen, or who experienced intolerable side-effects were given a test dose of ITB. Only those who had a satisfactory response were considered to be appropriate for pump implantation. All but one of the 152 patients had a satisfactory response, and the pump was implanted in 131 patients. Pre- and postoperative spasticity scores were compared and analyzed. The mean Ashworth score for rigidity decreased from 4.2 preoperatively to 1.3 (p < 0.0005) on ITB. The spasm score decreased from a mean of 3.4 to 0.6 (p < 0.0005). Reduction of spasticity resulted in improved levels of physical activity, decreased pain, and augmentation of sleep. Drug-related complications included constipation, muscular hypotonia, urinary retention, erectile dysfunction, nausea, dizziness, drowsiness, hypotension and bradycardia as well as tolerance to baclofen. Some patients experienced post-spinal puncture headaches. Catheter-related problems included occlusions, breaks, punctures, and dislodgments. Superficial pump pocket infection, pocket erosion, cerebrospinal fluid (CSF) leak, post-spinal puncture headache, and meningitis were some of the procedure-related complications. Two pumps flipped and another pump valve was stuck. We conclude that long-term intrathecal baclofen by an implanted programmable pump is a safe and effective method of treating severe intractable spinal spasticity.