S. A. Savage1, J. J. Sumislawski1, W. P. Dutton1, B. L. Zarzaur2 1University Of Tennessee Health Science Center Memphis,Memphis, TN, USA 2Indiana University-Purdue University Indianapolis,Indianapolis, IN, USA
Introduction: Retrospective studies of traumatic hemorrhage have embraced the concept of massive transfusion, in which all patients reaching a set volume (commonly ten units over 24 hours) are included. Patterns of hemorrhage vary, however, and massive transfusion definitions do little to describe these differences or their related outcomes. The purpose of this study is to describe subpopulations of hemorrhage in a cohort of injured patients from an urban Level One trauma center. We hypothesize that distinct group trajectories may be identified.
Methods: Patients requiring at least one unit of packed red blood cells (PRBC) in the first 24 hours of admission, and not suffering isolated head injury, were identified from June 2012 to May 2013. Time of blood transfusion, in minutes, was collected for each PRBC transfused in all patients. These times were then aggregated into 30-minute blocks over the first day. Group-based trajectory modeling was performed using Proc Traj (SAS, v. 9.3), with best model-fit determined using Bayesian Information Criterion (BIC) values.
Results: 318 patients met inclusion criteria for this study. 72% were male, the mean age was 39.5 years (SD 18), 39% suffered penetrating trauma, mean ISS was 17 (SD 11), mean 24 hour transfusion volume was 7 units PRBC (min 1u – max 50u) and overall mortality was 14%. Transfusion patterns for patients receiving > 10 units PRBC/24 hours (n=71) are shown in Figure 1. 12% of massive transfusion patients (group 1) actually received intermittent PRBC transfusions throughout the first day. All 318 patients were modeled to demonstrate 4 distinct trajectories for transfusion in figure 2. 20% of patients received negligible PRBC volumes over 24 hours (MIN- group1). 34% of patients received low but steady volumes of PRBC (LSV-group 2). 29% of patients received a moderate volume of PRBC only early in the hospital course (Early Bleeding (EB) – group 3). Patients in group 4 represent the massively bleeding subpopulation and comprise 17% of patients (MB). MB patients received large volumes of PRBC over the first seven hours and intermittently after that.
Conclusion: Traditional definitions of massive transfusion encompass both rapidly hemorrhaging patients, as well as those who are transfused gradually for other indications. These definitions are too broad. In this study, group-based trajectory modeling was used to demonstrate subpopulations of hemorrhage that are more clinically relevant. Understanding the trajectories of hemorrhage in injured subpopulations will allow more efficient allocation of clinical resources and concentrate research efforts on truly hemorrhaging subgroups.
