With the easier and earlier recognition of complex regional pain syndrome (CRPS), a reappraisal of its therapy, particularly the role and timing of sympathectomy, is warranted.Over a 9-year period, 42 patients with CRPS type II of the upper extremity were referred for sympathectomy. Patients were categorized according to the duration of the symptoms (group I, 3 months). All patients underwent initial medical treatment; stellate ganglion blocks were performed when symptoms persisted beyond 6 weeks. Patients were referred for thoracoscopic sympathectomy on persistence of the pain syndrome. A visual linear analogue scale was used to evaluate outcome of sympathectomy.Thoracoscopic dorsal sympathectomy was successfully undertaken in 32 patients. In the remaining 10 patients, thoracoscopy was not technically feasible and open sympathectomy was performed. There was an overall improvement in all 42 patients undergoing sympathectomy (P <.001, Wilcoxon signed rank test). The outcome in group I was significantly better than in group II (P <.003, Mann-Whitney U test). The diagnosis of sympathetically mediated pain with stellate blockade did not correlate with clinical outcome. Patients undergoing thoracoscopic sympathectomy had a better outcome than those undergoing open sympathectomy. There were no complications, and the hospital stay was shorter in the thoracoscopic group.Early recognition of CRPS and prompt recourse to surgical sympathectomy is a useful option in the management of CRPS.