UltraMTP in Adult Trauma Patients Undergoing Surgery Within 24 Hours: Effects on Mortality and Clinical Outcomes

The purpose of this study is to determine the effects of ultraMTP (>/=30 units pRBC within 24 hours) in trauma patients on mortality and secondary outcomes. The aim is to determine if there is a set number of pRBC units transfused in adult trauma patients undergoing surgery within 24 hours, after which mortality is inevitable and further transfusions are futile.

Rationale: To determine if there is a set number of pRBC units transfused in adult trauma patients undergoing surgery within 24 hours, after which mortality is inevitable and further transfusions are futile.

Intervention: Blood transfusions administered to adult trauma patients undergoing surgery within 24 hours of admission.

Objectives/Purpose: The main objective of this study is to examine the outcomes associated with large volume transfusions at U.S. level I trauma centers in adult trauma patients undergoing surgery within 24 hours of admission. The investigators aim to determine the mortality rate associated with ultraMTP (defined as >/= 30 units pRBC within 24 hours), and the effects of ultraMTP on secondary outcomes. The investigators would also like to compare outcomes among patients who receive various amounts of pRBC units within 24 hours. The information learned through this study will aid in determining whether there is a threshold transfusion amount, after which outcomes significantly worsen, and resuscitation efforts are futile and should cease, in order to save hospital resources, time, and costs.

Study Population/Sample Characteristics: Adult trauma patients requiring surgery within 24 hours of admission who receive blood products.

Study Methodology: This is a multicenter, retrospective observational study.

Study Endpoints/Outcomes:

• Primary outcome: 24-hour mortality (from the time of admission until 24 hours)
• Secondary outcomes: 1. ICU length of stay (LOS); 2. hospital LOS; 3. MV days; 4. Complications; 5. Multiple organ system failure; 6. in hospital mortality; 7. 30-day mortality; 8. discharge disposition.

Statistics/Analysis Plans:

The investigators will construct a logistic regression prediction model to 1) find the cut point classifications for pRBC units that optimize model prediction and 2) find additional variables that improve the predictive ability of our model. Previous studies on pRBC units and mortality have used different category cut points for pRBC units. The investigators will examine the functional form of pRBC units in its relationship with mortality to determine if there is a "plateau" effect of increasing pRBC units on mortality. The investigators will use this information to guide the classification of pRBC unit categories, comparing AIC values from several models with differing pRBC category classification cut off points to determine which of these are most highly associated with mortality. Using these categories, the investigators will then construct the prediction model, using the variables defined in Section 12. Interactions between these candidate predictors variables and pRBC units will additionally be considered. Model validation will be performed using k-folds cross-validation. Model performance will be evaluated by examining discriminative ability (area under ROC curve), calibration metrics, and optimism.