Plasma-First Resuscitation to Treat Haemorrhagic Shock During Emergency Ground Transportation in an Urban Area: A Randomised Trial
Authors: Moore HB, Moore EE, Chapman MP, et al.
Published in: Lancet, 2018 Jul 28; 392(10,144): 283–91.
When leaders and stakeholders use research to develop things like evidence-based guidelines, they must weigh that research’s quality. What they evaluate is often called strength of evidence or level of evidence. When making recommendations it is important to use the highest-level evidence available, and randomized trials are right at the top of the list.
In randomized trials subjects are assigned by chance to an experiment or control group. Randomization helps reduce bias, which is when results are influenced by factors unrelated to what’s being tested.
This month we review an interesting and important study funded by the U.S. Department of Defense. This study was part of the Control of Major Bleeding After Trauma trial, or COMBAT. COMBAT tested prehospital administration of plasma for control of major bleeding after injury. Its authors compared prehospital plasma administration to the standard of care, administration of normal saline.
This study was necessary because a 2016 review of the literature indicated a randomized, controlled trial was needed. Previously the military reported increased survival when plasma was administered during the Iraq war from 2003–2005. This was followed by retrospective studies that showed increased survival with early plasma administration.
However, some early clinical trials did not show any benefits to survival. Further, the authors indicated that prior to this study, the U.S. Food and Drug Administration was anticipated to approve lyophilized plasma, which prompted the Department of Defense to fund several randomized, controlled trials with the goal of producing robust evidence on early plasma resuscitation.
COMBAT was a pragmatic, randomized, placebo-controlled, single-center trial. A pragmatic trial is one designed to evaluate the effectiveness of interventions in real-world, routine-practice conditions. The trial took place within the paramedic division of Denver Health Medical Center from April 1, 2014 to March 31, 2017. The outcome of interest was mortality within 28 days. The authors tested the hypothesis that mortality would be lower among patients who received plasma before arrival at a Level 1 trauma facility than among those who received standard care with normal saline.
Consecutive hemorrhagic shock patients were randomized to receive either plasma or normal saline. Hemorrhagic shock was defined as systolic blood pressure less than or equal to 70 mm Hg or 71–90 mm Hg plus a heart rate of 108 bpm or more. Patients were excluded if they were less than 18 years old, were a prisoner, were pregnant, had an isolated gunshot wound to the head, were in asystole or had CPR performed before randomization, had a known objection to blood products, wore an opt-out bracelet or necklace, or if a family member objected to the patient’s enrollment. Once paramedics determined a patient was eligible, they began plasma or normal saline administration based on the contents of assigned coolers. Coolers were randomized to contain either plasma or frozen water by research coordinators. If the cooler contained frozen water, normal saline was administered based on hemodynamic need.
Prior to beginning the study, the authors performed a sample size calculation to determine the number of patients needed to detect a meaningful difference between the experiment and control groups. That calculation indicated 150 patients needed to be enrolled in the study. During the study period there were 144 potentially eligible patients identified and randomized to a group. Of these, 75 were initially assigned to the plasma group, and 69 were initially assigned to the saline control group.
However, 8 patients from the plasma group and 11 from the saline group met exclusion criteria. Recall the authors’ sample size calculation required them to enroll 150 patients. There’s an interesting reason why they didn’t get there, which we’ll return to shortly.
The authors chose to perform two statistical assessments. They performed an “as-treated” assessment that assessed patients based on whether they actually received plasma or normal saline. They also performed an “intent-to-treat” analysis. Intent-to-treat analyses are common in randomized trials. Essentially they analyze the subject based on the group to which they were assigned, whether or not they actually received plasma or normal saline. This is necessary in some trials due to noncompliance or missing data. This analysis generally gives a more conservative estimate of the treatment effect and helps researchers run statistical analyses without having to worry about protocol deviations, withdrawals, and anything else that could happen after randomization.
For example, in this study two patients were randomized to receive plasma but instead received normal saline because the paramedics misidentified the contents of the cooler. So the intent-to-treat analysis included the 75 patients initially assigned to the plasma group and the 69 assigned to the saline control group, while the as-treated assessment removed those who met exclusion criteria and included 65 patients in the plasma group and 60 in the control group. The results we will discuss in this review are from the as-treated analysis.
In the as-treated analysis, the plasma group included 52 (80%) males and 13 (20%) females. The saline control group included 51 (85%) men and 9 (15%) women. Median ages were 33 years for the plasma group and 32.5 years for the control group. The median transport time was 19 minutes for the plasma group and 16 minutes for the saline group (p=0.04).
All 65 patients in the plasma group received two full plasma units, with 21 (32%) receiving two units during transport, 24 (37%) receiving one unit during transport and a second in the emergency department, and 20 (31%) starting the first plasma unit during transport but, after its completion, having a second administered in the emergency department.
When comparing the two groups, the relative risk of death within 28 days was 1.54 with a 95% confidence interval of 0.60–3.98. Ten (15%) patients died within 28 days in the plasma group, and 6 (10%) in the control group (p=0.37).
Now, let’s get back to the sample size calculation that said this study needed to enroll 150 patients. Clinical trials are monitored by data safety monitoring boards. Essentially this board evaluates the results at specific intervals throughout the study to determine whether the study is safe for patients and should continue.
Throughout this study the board found no difference in mortality between the groups in either analysis. After 144 patients were enrolled, the data safety monitoring board and FDA deemed the study futile and approved its termination because “the outcome had not differed in any of the interim analyses, indicating that no difference should be anticipated.”
It should be noted that this comparison occurred in an urban environment with relatively short transport times to a Level 1 trauma center. The authors stated that while no improvement in clinical outcomes was found, no adverse events were found either. They indicated further study is necessary because the use of plasma might be beneficial in rural or wilderness environments with longer transport times.
As with all studies, this study had some limitations. The authors noted that some hypotension may not have been due to hemorrhage, which may account for the lack of effect seen in the plasma group. Further, assignment of patients to the control group may have freed paramedics from performing additional tasks related to plasma administration, allowing them to better focus on patient care. Finally, this study took place in a system that had plasma immediately available upon arrival at the hospital. Results may differ in other systems.
Antonio R. Fernandez, PhD, NRP, FAHA, is research director at the EMS Performance Improvement Center and an assistant professor in the Department of Emergency Medicine at the University of North Carolina–Chapel Hill.