2.1. Development of the Scoring System

Data of trauma patients admitted to the Niguarda Hospital were prospectively collected between February 2016 and December 2018. Inclusion criteria were: (i.) age ≥ 18 years, (ii.) Hospital Trauma Team activation by the emergency call center, and (iii.) at least one episode of systolic blood pressure (SBP) < 90 mmHg during PH setting or upon arrival in the ED. Exclusion criteria were: (i.) isolated brain (TBI) or spinal cord injury to avoid potential hemodynamic bias, (ii.) confirmed pregnancy, (iii.) third degree burns in any body region, (iv.) secondary referral from another hospital, (v.) death within the first 30 min of ED arrival and (vi.) ED arrival > 120 min after trauma. Demographics, Injury Severity Score (ISS), type of trauma (blunt or penetrating), PH and ED vital parameters (GCS, SBP, HR, Revised Trauma Score, RTS), PH tranexamic acid (TXA) administration, positive abdominal extended focused abdominal sonography for trauma (E-FAST), pelvic fracture, hemothorax, limb amputation, blood lactate and base excess (BE) levels on admission and standard coagulation parameters (e.g., PT, PTT, platelets count) were evaluated.

Hemodynamically unstable patients as defined by the Advanced Trauma Life Support (ATLS^®) criteria of classes III or IV at the scene were treated with one gram bolus of TXA by the PH personnel. In non-responder patients, (SBP persistently < 90 mmHg notwithstanding repeated 250 mL crystalloid bolus infusions), hypotesizing a critical bleeding, MT (≥4 blood units within the first 60 min from ED admittance) is foreseen. Massive transfusion protocol was administered as follow: - two 0-negative packed red blood cell promptly available on ED, are transfused, followed by cross-matched packed red blood cells (PRBCs)—fresh frozen plasma (FFP)—platelets (PLT) at PRBCs:FFP:PLT 1:1:1 ratio. In our hospital the average volume of a PRBC is 250–300 mL per bag, obtained by removing 200–250 mL of plasma from 450–500 mL of whole blood.

In addition, cryoprecipitate 1 unit/10 kg body weight if fibrinogen < 2 gr/L in order to correct the consumption of coagulation factors and TXA 1 gr infusion over 8 h, to avoid or counteract the fibrinolysis are administered.

Crystalloid infusion was reduced to a minimum and colloids avoided during both PH and in-hospital care. This protocol remained unchanged throughout the study period. The source of bleeding was addressed and identified in the ED through chest X-ray, pelvis X-ray and E-FAST. A contrast-enhanced whole-body CT-scan (WBCT) was carried out only after hemodynamic stabilization. All DCS techniques were applied if indicated, including damage control laparotomy, extra- peritoneal packing, emergency thoracotomy, limb amputation, external long bone and pelvis fixation. Angioembolization was performed if WBCT after surgery indicated persistent bleeding. Resuscitative endovascular balloon occlusion of the aorta (REBOA) was not available during the study period.

In patients with normalizing SBP over 90 mm Hg after initial crystalloids infusion, MT protocol was not activated and blood transfusions were eventually driven when required by hemoglobin levels.

Patients were divided into those who had critical bleeding and those who did not. MT was defined as 4 or more blood units during the first hour of admission to the ED (MT+) [9]. Those patients with absent signs of critical bleeding and no massive transfusion protocol initiated were defined as MT−. Statistical analysis was performed using the IBM SPSS Statistics version 21 Software. The continuous variables were expressed as mean with standard deviations (M ± SD) and median and compared using the Student’s t test. The categorical variables were compared by chi-square test. A p-value < 0.05 was considered statistically significant. For each potential predictive variable for MT a logistic model was generated to obtain an odds ratio (OR) with relative 95% confidence intervals to investigate relationships between variables and MT. Clinically relevant variables were entered into multivariable logistic regression models to identify independent predictors of MT and to derive a score for early clinical decision-making.