The objective of this study is to undertake a population based study on the incidence, prevalence, natural history, and response to treatment of complex regional pain syndrome (CRPS). All Mayo Clinic and Olmsted Medical Group medical records with codes for reflex sympathetic dystrophy (RSD), CRPS, and compatible diagnoses in the period 1989-1999 were reviewed as part of the Rochester Epidemiology Project. We used IASP criteria for CRPS. The study population was in the Olmsted County, Minnesota (1990 population, 106,470). The main outcome measures were CRPS I incidence, prevalence, and outcome. Seventy-four cases of CRPS I were identified, resulting in an incidence rate of 5.46 per 100,000 person years at risk, and a period prevalence of 20.57 per 100,000. Female:male ratio was 4:1, with a median age of 46 years at onset. Upper limb was affected twice as commonly as lower limb. All cases reported an antecedent event and fracture was the most common trigger (46%). Excellent concordance was found between symptoms and signs and vasomotor symptoms were the most commonly present. Three phase bone scan and autonomic testing diagnosed the condition in >80% of cases. Seventy-four percent of patients underwent resolution, often spontaneously. CRPS I is of low prevalence, more commonly affects women than men, the upper more than the lower extremity, and three out of four cases undergo resolution. These results suggest that invasive treatment of CRPS may not be warranted in the majority of cases.