Although emergency department thoracotomy (EDT) is often considered a controversial "last chance" method of resuscitation, we hypothesized that EDT performed in a busy urban Level I trauma center has significant salvage rates despite the absence of traditional survival predictors. A retrospective review revealed that 180 patients underwent EDT after traumatic arrest for penetrating injury between 2000 and 2005. All were deemed nonsalvageable by other resuscitation methods. Injury mechanism and location, signs of life (SOLs), initial cardiac rhythm, and presence of vital signs were analyzed. In total, 23 patients survived hospitalization neurologically intact. Compared to nonsurvivors, survivors more often suffered multiple stab wounds (21.7% vs. 1.9%, p = 0.001), presented with field (95.7% vs. 72.6%, p = 0.016) and ED (87.0% vs. 60.5%, p = 0.014) SOLs, had sustainable cardiac rhythms (sinus tachycardia, 43.5% vs. 10.2%, p = 0.001; normal sinus rhythm, 17.4% vs. 4.5%, p = 0.037), and had measurable vital signs (65.2% vs. 25.5%; p = 0.001). However, only 3 of 23 (13.0%) survivors had all survival predictors, and one survivor had none. Frequent predictors in survivors were field SOLs (95.7%), ED SOLs (87.0%), salvageable initial cardiac rhythms (78.3%), and obtainable vital signs (65.2%). Stabbing mechanism (30.4%) and cardiac injury location (30.4%) were least common. Had a strict policy of EDT performance based solely on the presence of survival predictors been followed and EDT withheld, several patients who ultimately survived would have died. Our study suggests that EDT is a technique that should be utilized for patients with critical penetrating injuries even in the absence of many traditional survival predictors.