Predictors of outcome in 215 patients requiring emergency department thoracotomy: A 2-year prospective study in a large urban trauma center.

Abstract

Background: Emergency department thoracotomy is a formidable procedure. We prospectively evaluated physiologic parameters in patients undergoing emergency department thoracotomy, mortality at the scene of the traumatic incident, and upon arrival, including cardiopulmonary resuscitation time, cardiovascular respiratory score, time to emergency department thoracotomy, and restoration of blood pressure, and identified predictors of outcomes.

Methods: This was a prospective 2-year study of patients in cardiopulmonary arrest requiring emergency department thoracotomy in a trauma center. Statistical analysis included univariate and stepwise logistic regression analysis.

Results: Two hundred fifteen patients admitted in cardiopulmonary arrest required emergency department thoracotomy. Penetrating injury (n = 167, 78%), gunshot wounds (n = 142, 78.0%), stab wounds (n = 21, 12.6%), shotgun wounds (n = 4, 2.4%), and blunt (n = 48, 22%). The mean revised trauma score was -0.6, the mean injury severity score was -42, and the mean duration of cardiopulmonary resuscitation was 12 minutes. Emergency department deaths totaled 162 (75%), and 1-hour survivors totaled 53 (25%). Overall survivors were 6 of 215 (2.8%). Penetrating cardiac injury survivors were 6 of 74 (8.2%). Two groups were compared: (1) emergency department deaths versus 1-hour survivors-all physiologic parameters were predictive of outcomes (P < .0001) and (2) 1-hour survivors versus overall survivors-no physiologic parameters were predictive of outcome. Duration of cardiopulmonary resuscitation (P = .04), penetrating mechanism (P < .001), and exsanguination (P < .006) were significant. The cardiovascular respiratory score revealed a trend (P = .007). Death versus overall survival-restoration of blood pressure was predictive (P < .006). Stepwise logistic regression, using a 4-factor model, including no spontaneous ventilation, absence of pulse upon arrival, injury severity score >45, and blunt mechanism of injury, was 99.3% predictive of mortality.

Conclusion: For patients presenting in arrest requiring emergency department thoracotomy, all parameters plus a cardiopulmonary resuscitation duration of 12 minutes are predictive of mortality. Seventy-five percent of patients can be excluded from emergency department thoracotomy. Emergency department thoracotomy should be applied selectively to patients sustaining penetrating cardiac injuries and some vascular injuries.