Twenty-five patients suffering from intractable pain due to a chronic spinal cord lesion underwent a percutaneous test of spinal cord stimulation. At the end of the test period, 40.9% of the patients reported a mean of 65% pain relief and these patients were selected for ongoing stimulation. At a mean follow-up time of 37.2 months the success rate, based on the number of patients with more than 50% pain relief, had fallen to 18.2%. Pain relief rates were analyzed in relation to quality of pain, neurological status, level and extent of the lesion, and electrode level to identify prognostic factors that could improve the clinical usefulness of spinal cord stimulation. Patients experiencing painful spasms or a constrictive type of pain and with incomplete thoracic lesions were found to be the best candidates for spinal cord stimulation.