This study was designed to examine the results of emergency room resuscitative thoracotomy (ERRT) and to formulate cost-effective indications for this procedure. A retrospective study was performed of 28 patients who had ERRT at St. Elizabeth Hospital Medical Center, Youngstown, Ohio, during the 4 years from July 1985 through June 1989. The prognostic factors analyzed included mechanism and site of injury, signs of life (SOL), vital signs (VS), age, gender, and prehospital care. The overall survival rate of ERRT was 7% (2 of 28 patients). The survival rate was 18% (2 of 11 patients) with penetrating trauma, and 0% (none of 17 patients) with blunt trauma. The best survival rate was 66% in the subgroup of patients with penetrating trauma and SOL present at the scene and in the emergency room (ER), (two of three patients). Our observations were combined with those of 23 studies from the literature involving 2294 trauma patients who had ERRT. Using meta-analysis, the survival rate was 11% overall. Improved survival was noted for patients with penetrating trauma compared with patients with blunt trauma (14% vs. 2%, p < 0.01). There were no survivors in the group of patients with no SOL at the scene, and there were no neurologically intact survivors among blunt trauma patients with no SOL upon arrival at the ER. An algorithm based on mechanism of injury and presence or absence of SOL at the scene and in the ER is proposed. This algorithm would decrease the number of ERRTs performed by 41% without decreasing the number of neurologically intact survivors.(ABSTRACT TRUNCATED AT 250 WORDS)