Limited research suggests that there may be Warm complex regional pain syndrome (CRPS) and Cold CRPS subtypes, with inflammatory mechanisms contributing most strongly to the former. This study for the first time used an unbiased statistical pattern recognition technique to evaluate whether distinct Warm vs Cold CRPS subtypes can be discerned in the clinical population. An international, multisite study was conducted using standardized procedures to evaluate signs and symptoms in 152 patients with clinical CRPS at baseline, with 3-month follow-up evaluations in 112 of these patients. Two-step cluster analysis using automated cluster selection identified a 2-cluster solution as optimal. Results revealed a Warm CRPS patient cluster characterized by a warm, red, edematous, and sweaty extremity and a Cold CRPS patient cluster characterized by a cold, blue, and less edematous extremity. Median pain duration was significantly (P < 0.001) shorter in the Warm CRPS (4.7 months) than in the Cold CRPS subtype (20 months), with pain intensity comparable. A derived total inflammatory score was significantly (P < 0.001) elevated in the Warm CRPS group (compared with Cold CRPS) at baseline but diminished significantly (P < 0.001) over the follow-up period, whereas this score did not diminish in the Cold CRPS group (time × subtype interaction: P < 0.001). Results support the existence of a Warm CRPS subtype common in patients with acute (<6 months) CRPS and a relatively distinct Cold CRPS subtype most common in chronic CRPS. The pattern of clinical features suggests that inflammatory mechanisms contribute most prominently to the Warm CRPS subtype but that these mechanisms diminish substantially during the first year postinjury.