Background: Exsanguination as a syndrome is ill defined. The objectives of this study were to investigate the relationship between survival and patient characteristics—vital signs, factors relating to injury and treatment; determine if threshold levels of pH, temperature, and highest estimated blood loss can predict survival; and identify predictive factors for survival and to initiate damage control.

Material and methods: A retrospective 6-year study was conducted, 1993 to 1998. In all, 548 patients met one or more criteria: (1) estimated blood loss ≥2,000 mL during trauma operation; (2) required ≥1,500 mL packed red blood cells (PRBC) during resuscitation; or (3) diagnosis of exsanguination. Analysis was made in two phases: (1) death versus survival in emergency department (ED); (2) death versus survival in operating room (OR). Statistical methods were Fisher’s exact test, Student’s t test, and logistic regression.

Results: For 548 patients, mean Revised Trauma Score 4.38, mean Injury Severity Score 32. Penetrating injuries 82% versus blunt injuries 18%. Vital statistics in emergency department: mean blood pressure 63 mm Hg, heart rate 78 beats per minute. Mean OR pH 7.15 and temperature 34.3°C. Mortality was 379 of 548 (69%). Predictive factors for mortality (means): pH ≤7.2, temperature <34°C, OR blood replacement >4,000 mL, total OR fluid replacement >10,000 mL, estimated blood loss >15 mL/minute (P <0.001). Analysis 1: death versus survival in ED, logistic regression. Independent risk factors for survival: penetrating trauma, spontaneous ventilation, and no ED thoracotomy (P <0.001; probability of survival 0.99613). Analysis 2: death versus survival in OR, logistic regression. Independent risk factors for survival: ISS ≤20, spontaneous ventilation in ED, OR PRBC replacement <4,000 mL, no ED or OR thoracotomy, absence of abdominal vascular injury (P <0.001, max R² 0.55, concordance 89%).

Conclusions: Survival rates can be predicted in exsanguinating patients. “Damage control” should be performed using these criteria. Knowledge of these patterns can be valuable in treatment selection.