Traumatic injury is a leading cause of death throughout the world, and subsequent uncontrolled hemorrhage is the killer. These deaths are often preventable with modern hemorrhage control techniques and resuscitation approaches derived from combat (Damage Control Resuscitation). During the wars in Iraq and Afghanistan, retrospective data suggested that a novel – yet archaic – approach to resuscitation with balanced transfusions approximating whole blood improved survival, but this data was retrospective and included many confounders and sometimes conflicting results. Better data were needed. This prompted the PROMMTT study, an observational study where research assistants recorded transfusion practices in traumatically injured patients in ten separate trauma centers. The PROMMTT results were enlightening and paved the way for a more formal prospective, randomized, pragmatic trial – PROPPR. These two studies and their data inform current trauma resuscitation strategies, and confirmed and codified a more balanced approach for future practice. These two articles are reviewed below and their nuances explored.
The PROMMTT Study, published in JAMA Surgery in 2013, was a multi-center, prospective, observational study that “described when RBCs, plasma, and platelets were infused and assessed the association between in-hospital mortality and the timing and amount of blood products.”
DISCUSSION: PROMMTT was a simple yet fascinating study that generated numerous interesting data and revealed the standard of US trauma care circa 2010. Ten US Level 1 trauma centers participated in the data collection where on-call research assistants documented the trauma team transfusion practices in real time for patients that met the study inclusion criteria. Personally, I would love to see this study repeated a decade later to see how trauma transfusion practices have changed. A few salient points are discussed below.
Mortality. Mortality was high, one of “the highest of any acute surgical disease process.” 25% of patients in the analysis cohort died, and those who died of hemorrhage died early. 58% of hemorrhagic deaths occurred within 3 hours of arrival, and 94% of hemorrhagic deaths occurred within the first 24 hours. The median time for hemorrhagic death was 2.6 hours. Traumatic hemorrhage kills often and kills early. So, if you want to affect mortality, you must act quickly.
Higher Ratios. The PROMMTT investigators, due to practice variability and for clinical utility, divided the patients into groups of low ratio (<1:2), moderate ratio (>1:2 but <1:1), and high ratio (>1:1) with respect to the unit ratios of plasma:RBCs and platelets:RBCs transfused. Then they compared these ratios to multiple time intervals spanning from 30 minutes after ED arrival to 30 days later. A high ratio (1:1) of plasma or platelets to RBCs means that the patients received approximately equivalent units of blood products, which most approximates whole blood. If a patient received a low ratio (<1:2), it means that the patient was transfused more units of RBCs than units of plasma or platelets (old school approach). The PROMMTT data showed that there was a “protective association between higher transfusion ratios and mortality” within the first 6 hours, but this association diminished after that initial time frame.
Variability. Unfortunately, transfusion practices varied widely amongst the participating trauma centers, and plasma and platelets were often delayed. Much of the trauma literature coming out of the wars in Iraq and Afghanistan at that time emphasized the necessity of balanced transfusions, but these data were far from robust (most retrospective) and this practice had not yet widely translated to the civilian sector. In the PROMMTT study, at 30 minutes, only a third of patients had received plasma and only 1% had received platelets. At three hours, 90% of patients had received plasma, and two-thirds of patients had received platelets (see the table below).
The PROPPR Trial, published in JAMA in 2015, was a multi-center, pragmatic, prospective, randomized, controlled clinical trial that compared mortality in traumatically injured patients who were transfused blood products in a ratio of 1:1:1 versus 1:1:2.
DISCUSSION: The PROPPR Trial was a very well performed study whose design was structured from the evidence gathered in the PROMMTT Study. It met most of the qualifications required of high quality trials: it included multiple large academic centers, it was prospective, randomized, and controlled. Blinding was impossible practically, but the trial was designed to mitigate potential non-blinded biases as much as possible. The treating physicians were unaware of the randomization group until the blood product containers were opened in the trauma bay, and those clinicians determining cause of death later were separate from that site and blinded to the treatment group.
Mortality: While this trial did not detect a statistically significant difference in mortality at 24 hours and 30 days, it did demonstrate a difference. The Kaplan-Meier Curves comparing the probability of death and time to death best demonstrate this disparity. The trial size (680) was originally powered to detect a 10% mortality difference at 24 hours and a 12% difference at 30 days (a difference the authors considered to be clinically significant). The discovered mortality difference of 4.2% at 24 hours would have required a sample size of 2,968 to potentially achieve statistical significance, which would have been a nearly impossible sample size to achieve logistically. This trial was conducted at 12 different trauma centers over the course of 17 months, and 14,313 high-level traumas were activated at these centers during that time. Approximately 80% of the those traumas were screened for the trial, 10,505 were excluded, and 680 were randomized into the study. This was a massive undertaking and would be difficult to replicate practically at a much larger scale.
Exsanguination and Hemostasis: While the authors did not find a difference in overall mortality in the two groups, fewer died from exsanguination in the 1:1:1 group (31 vs. 50, -5.4% difference (CI = -10.4 to -0.5)) and more achieved hemostasis in the 1:1:1 group (291 vs. 267, p=0.006). The 24 hour and 30 day mortality markers are completely arbitrary timeframes, representing only the hours in a day and days in a month, and do not represent any significant physiologic parameter. Death from exsanguination occurred at a median of 2.6 hours in PROMMTT, and at a median of 2.3 hours in PROPPR, long before either of the primary outcomes occurred. And beyond the acute time frame, and especially after 24 hours, deaths from traumatic brain injury predominated likely further confusing the mortality picture.
1:1:1 versus 1:1:2: The ratios investigated in this trial may be the largest point of contention for some, though these ratios were driven by the evidence from the PROMMTT Study. The PROMMTT study demonstrated leading edge trauma resuscitation at large academic centers in 2010 where some trauma teams had already begun to adopt the mindset of resuscitating with blood product ratios that nearly approximated whole blood. While the transfusion practices varied widely in PROMMTT, the resuscitations more closely approached 1:1:1 than was common decades earlier. So while the ratios chosen for the PROPPR study were evidence based and represented 2010 practice, they may have been too similar to find a large mortality difference. I believe that if the ratios representing trauma resuscitation in the early 2000s or earlier were compared to 1:1:1, that a large and clinically significant difference in mortality would have been demonstrated. Although by 2010, trauma resuscitation had evolved and transfusing at lower ratios would have been ethically unsound.
Adverse Outcomes: Concerns over potential adverse outcomes (“acute respiratory distress syndrome, multiple organ failure, infection, venous thromboembolism, and sepsis”) drove some clinicians to avoid the 1:1:1 resuscitation strategy, but the PROPPR trial demonstrated no significant increase in these outcomes in the 1:1:1 group despite the higher volumes of transfused plasma and platelets. Therefore resuscitating with a 1:1:1 strategy is likely safe.
The PROMMTT Study showed that death from exsanguination in trauma occurs early (median 2.6 hours), and that transfusion ratios closely approximating whole blood (1:1:1) were associated with increased survival within 6 hours of injury. The PROPPR Study demonstrated that death from exsanguination was decreased and that achieving hemostasis was more likely in the first 24 hours when utilizing a transfusion strategy that approximates whole blood (1:1:1).
When resuscitating the severely injured trauma patient, transfuse and balance the blood products as early as possible. We should forsake any previous practice of initially transfusing multiple units of RBCs followed by plasma and platelets later as dictated by labs. The critically traumatically injured will die long before receiving appropriately balanced resuscitation products using that archaic approach. Interestingly, many trauma centers have progressed to transfusing plasma and platelets first before RBCs, further studies are investigating the utility of this approach, but anecdotally this seems the way of the future and may physiologically relate to the reversal of coagulopathy. Finally, embrace the balanced transfusion strategy demonstrated by PROMMTT, PROPPR, and their preceding retrospective studies from the point of injury until hemostasis to offer the patient the best chance of survival.
