Epidural spinal cord stimulation by means of chronically implanted electrodes was carried out on 121 patients with pain of varied benign organic etiology. In 116 patients, the pain was confined to the back and lower extremities and, of these, 56 exhibited the failed-back syndrome. Most patients were referred by a pain management service because of failure of conventional pain treatment modalities. Electrodes were implanted at varying sites, dictated by the location of pain. A total of 140 epidural implants were used: 76 unipolar, 46 Resume electrodes, 12 bipolar, and six quadripolar. Patients were followed for periods ranging from 6 months to 10 years, with a mean follow-up period of 40 months. Forty-eight patients (40%) were able to control their pain by neurostimulation alone. A further 14 patients (12%), in addition to following a regular stimulation program, needed occasional analgesic supplements to achieve 50% or more relief of the prestimulation pain. Pain secondary to arachnoiditis or perineural fibrosis following multiple intervertebral disc operations, when predominantly confined to one lower extremity, seemed to respond favorably to this treatment. Uniformly good results were also obtained in lower-extremity pain secondary to multiple sclerosis. Pain due to advanced peripheral vascular disease of the lower limbs was well controlled, and amputation below the knee was delayed for up to 2 years in some patients. Pain due to cauda equina injury, paraplegic pain, phantom-limb pain, pure midline back pain without radiculopathy, or pain due to primary bone or joint disease seemed to respond less well. Patients who responded to preliminary transcutaneous electrical nerve stimulation generally did well with electrode implants. Notable complications included wound infection, electrode displacement or fracturing, and fibrosis at the stimulating tip of the electrode. Three patients in this series died due to unrelated causes. Epidural spinal cord stimulation has proven to be an effective and safe means of controlling pain on a long-term basis in selected groups of patients. The mechanism of action of stimulation-produced analgesia remains unclear; further studies to elucidate it might allow spinal cord stimulation to be exploited more effectively in disorders that are currently refractory to this treatment modality.