This study includes our 12-year experience with chemical sympathetic blocks and surgical sympathectomies for causalgic pain of reflex sympathetic dystrophy (RSD) with emphasis on factors affecting clinical outcome. Medical records of patients undergoing sympathectomies for causalgic pain were analyzed. The patients were classified according to Drucker et al. as stage I, II, or III. Results of chemical and surgical sympathectomies were analyzed using both univariate and multivariate methods. Twenty-one patients had lumbar and seven had cervicodorsal sympathectomies for RSD. The mean duration between initial injury and chemical sympathetic block was 10 months with a mean of 11.4 months to surgical sympathectomy. Ten patients (36%) had overt extremity trauma as the precipitating event. Ten patients (36%) had a lumbar laminectomy, three of whom developed the syndrome bilaterally. There was no operative mortality; however, 25% had transient postoperative sympathetic neuralgia. The early and late (> 6 months) satisfactory outcomes after surgical sympathectomy were 82% and 71%, respectively. Patients with stage II presentations were significantly more likely to have satisfactory early (92%) and late (79%) outcomes than stage III patients, 0% and 0% (p = 0.019). Patients with an excellent response to chemical sympathetic block were more likely to have satisfactory early and late surgical outcomes. The time between injury and chemical block and surgical sympathectomy was significantly shorter in patients who had satisfactory early and late surgical outcomes (p < 0.0001). Multivariate analyses demonstrated that the most important independent factor in determining early and late satisfactory outcomes of sympathectomy was the time between injury and sympathectomy (p = 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)