Red, White and Grey: Is blood truly superior?

A synopsis of the RePHILL trial and examination of current clinical guidelines

· Clinical Education,Research

 Words by Ali Rengers

Ask any clinician their thoughts on tissue reperfusion with crystalloids and you’ve opened a can of worms. Answers range from the black and white – “I would rather keep an actively bleeding patient’s systolic BP (SBP) at 60mmHg with no fluids than give a litre of crystalloid”1 to the grey ‘well… it depends’.

The recent trial by Crombie et al., compared the use of packed red blood cells (PRBC) and lyophilised (dried by freezing) plasma (LyoPlas) to 0.9% sodium chloride in improving tissue reperfusion and reducing mortality in the prehospital setting. The aim of their trial entitled “Resuscitation with blood products in patients with trauma-related haemorrhagic shock receiving prehospital care (RePHILL): a multicentre, open-label, randomised, controlled, phase 3 trial” was to determine if the blood products were in fact superior to the crystalloid in treating prehospital haemorrhagic shock. The study aimed to determine this by measuring mortality (death between injury and discharge from the primary receiving facility) and blood capillary lactate clearance (failure of this clearance was defined as less than 20% in the first two hours after the randomisation process began). Lactate in this instance is not lactic acid, but blood lactate; the waste product released from oxygen deprived tissues that requires recycling or oxidisation through the liver or other body tissues2.

The current preferred choice for prehospital reperfusion of patients experiencing traumatic haemorrhage is with PRBC 3. According to Griggs et al., the administration of PRBC to civilian patients stemmed from its successful battlefield use in Afghanistan and Iraq 3. The goal of battlefield administered PRBC to traumatic haemorrhage patients was to restore haemoglobin concentration and moderate the possibility of acute traumatic coagulopathy4. This type of coagulopathy is characterised by the patient’s initial inability to clot, followed by excessive clotting causing an increased risk of thrombosis, and is one element of the ‘lethal triad’ of haemorrhagic shock4, 5.

According to Crombie et al., although the use of PRBC and plasma in the prehospital environment seems intuitive, it has its drawbacks. Sustaining universal blood products, managing blood product wastage and compliance, safe and timely administration and identification of suitable patients in the prehospital setting all need consideration.

Based across four prehospital critical care services in the UK the study was a multicentre, phase three, allocation concealed, open-label, parallel group, randomised controlled trial. Now while this all seems very complex, broken into small chunks it is quite understandable. Phase three denotes that the trial was undertaken to study the efficacy of the intervention compared to other standard or experimental interventions with a large group of individuals6. Allocation concealed involves separating the individual involved in recruiting patients from the randomisation process7. Open label, also referred to as unblinded trial, allows the participants and researchers involved to know the treatments being used, while parallel group involves groups of patients observed for the same time after receiving different treatments. Finally, a randomised controlled trial involves participants being randomly distributed into groups with the non-control participants receiving the intervention and the control group receiving another treatment or placebo. So, the study was of a reasonable sample size with unbiased randomisation. The participants and researchers knew what the treatments being used were, however did not know who received which treatment over the same observation time. Essentially, a gold standard study.

Participants were included in the study if they were older than 16 years with a traumatic injury and hypotension (SBP less than 90mmHg or the absence of a radial pulse). Participants were excluded if they received prehospital blood products before being assessed for trial eligibility, were known to refuse prehospital blood products, were pregnant, had an isolated head injury without major haemorrhage, were prison inmates, or patients in traumatic cardiac arrest prior to arrival of the prehospital emergency team. A total of 432 participants were involved in the study with 209 randomly assigned to the blood product intervention group and 223 to the 0.9% sodium chloride group. Written informed consent was sought from the patient or their personal or professional representative after the emergency had passed.

The randomisation of PRBC/LyoPlas and 0.9% sodium chloride involved placing either product into sealed, externally identical, temperature monitored boxes. These boxes were then carried on emergency vehicles for up to 48 hours before replacement if unused. Blood products were then returned to the blood bank stock.

Intervention group participants were supplied with up to two units of PRBC and two units of LyoPlas during the trial. Boluses between 220mL – 340mL (mean = 282mL) of group O, rhesus factor D negative (the protein found on the outside of red blood cells (RBC)), Kell negative (the antigen attached to the membrane of RBC) leucodepleted (removal of granulocytes and lymphocytes) PRBC in an additive solution were administered to participants. LyoPlas (derived from blood group A or AB) was reconstituted in 200mL of water for injection (total bolus of 213mL) immediately prior to administration. One unit of PRBC was alternated with one unit of LyoPlas when administering the intervention to patients. The control group received a maximum of four, 250mL 0.9% sodium chloride boluses.

In both instances fluid delivery was administered with a fluid warmer, and interventions were dispensed until arrival at hospital or return of palpable radial pulse or a SBP of 90mmHg or greater. If treatment was ceased for any of these reasons and the participant’s SBP decreased on the way to hospital, fluid administration was reinstated. If intervention participants received all four units of blood products and required further fluid replacement, non-trial 0.9% sodium chloride was administered.

Participants were mostly caucasian (62%) with a median age of 38 years. Road traffic collision accounted for 62% of major trauma. All patients were administered 430mL of crystalloid fluid and 90% tranexamic acid (TXA) prior to randomisation. Of the patients receiving PRBC/LyoPlas, 64% failed to clear lactate or died or both, while 65% of participants administered 0.9% sodium chloride also reached this primary outcome. Although the mean haemoglobin concentration was higher in patients receiving PRBC/LyoPlas on emergency department (ED) arrival, lactate concentrations, vital signs and 3 hour and 30-day mortality rates were similar across the PRBC/LyoPlas and 0.9% sodium chloride groups.

Patients receiving PRBC/LyoPlas experienced higher rates of respiratory, neurological, cardiovascular and liver failure when compared to their 0.9% sodium chloride receiving counterparts (58% vs. 52%; 63% vs. 57%; 67% vs. 62%; 9% vs. 5%). They also experienced higher rates of acute respiratory distress syndrome (ARDS) compared to the 0.9% sodium chloride group (6% vs. 2%). Failure to coagulate was higher in patient’s receiving 0.9% sodium chloride than those receiving blood products (15% vs. 8%).

Infection rates were similar across both groups, however patients receiving 0.9% sodium chloride experienced higher levels of urinary tract infections (UTI) (8% vs. 4%), while PRBC/LyoPlas patients faced a higher incidence of soft tissue infection (25% vs. 15%). Transfusion related complications in the first 24 hours were similar across both groups (7%).

The use of blood products was overall higher in patients receiving PRBC/LyoPlas therapy. Based on the trial findings highlighted above, the monetary and logistical costs of prehospital blood product use may not be justified. Current European guidelines recommend that commencement of transfusion should occur when haemoglobin levels fall between 70-90g/L. Patients that received PRBC/LyoPlas were delivered to the ED with higher concentrations of haemoglobin (133g/L vs. 118g/L), however only 6% of individuals receiving 0.9% sodium chloride arrived at the ED with haemoglobin levels less than 80g/L.

Crombie et al. believe that the lack of demonstrated benefit of blood products may be due to several factors, including short patient transportation times to definitive care, use of lypophilised rather than fresh frozen plasma and enrolment of predominantly older participants in the study with higher injury severity.

Overall, the RePHILL trial did not demonstrate the superiority of PRBC/LyoPlas therapy over 0.9% sodium chloride in improving patient mortality or lactate clearance in prehospital haemorrhagic shock.

The RePHILL trial could have implications for major services nationally and internationally. According to the Queensland Ambulance Service (QAS) guidelines “early blood use is preferrable in haemorrhagic shock”8. Available blood products are to be administered using a warmer at a 1:1 ratio of PRBC to plasma 8, which sounds like an involved process in the prehospital setting. The New South Wales (NSW) Institute of Trauma and Injury Management state that the consensus believe the “early use of blood, if available, remains the optimal resuscitation fluid for the hypovolaemic patient”9.

According to the RePHILL trial administration of blood products to patients experiencing traumatic haemorrhage in the prehospital setting does not improve patient mortality. Although further studies will need to be conducted to corroborate the findings of the RePHILL trial, the results are exciting and could influence evidence-based practice guidelines for prehospital treatment of haemorrhagic shock in the not-too-distant future.

The original article can be read here.


1. @karimbrohi. “I would rather keep a BLEEDING patient’s systolic BP at 60mmHg with no fluids than give a litre of crystalloid”. [10 Jan 2019.] [Accessed 28 Mar 2022]

2. Levitt DG, Levitt JE, Levitt MD. Quantitative assessment of blood lactate in shock: Measure of hypoxia or beneficial energy sources. Biomed Res Int [Internet]. 2020 Oct [cited 2022 Jun 21]; 2020:1-24. Available from:

3. Griggs JE, Jeyanathan J, Joy M, Russell MQ, Durge N, Bootland D, et al. Mortality of civilian patients with suspected traumatic haemorrhage receiving pre-hospital transfusion of packed red blood cells compared to pre-hospital crystalloid. Scand J Trauma Resusc Emerg Med [Internet]. 2018 Nov [cited 2022 Jun 21]; 26:1-9. Available from:

4. Penn-Barwell JG, Roberts SA, Midwinter MJ, Bishop JR. Improved survival in UK combat casualties from Iraq and Afghanistan 2003-2012. J Trauma Acute Care Surg [Internet]. 2015 May [cited 2022 Jun 21]; 78(5):1014-1020. Available from:

5. Dawes R, Thomas GO. Battlefield resuscitation. Curr Opin Crit Care [Internet]. 2009 Dec [cited 2022 Jun];15(6):527-535. Available from:

6. Phases of clinical trials [Internet]. Australian Government National Health and Medical Research Council Department of Industry, Innovation and Science; 2015 [cited 2022 Mar 27]. Available from:

7. Clark L, Fairhurst C, Torgerson, DJ. Allocation concealment in randomised controlled trials: are we getting better? [Internet]. BMJ. 2016;355(8083):i5663. doi:10.1136/bmj.i5663

8. Queensland Ambulance Service. Clinical practice guidelines: Other/ haemorrhage control. [Internet]. Queensland; Queensland Ambulance Service; 2021.

9. New South Wales Institute of Trauma and Injury Management. Adult trauma clinical practice guidelines: Management of hypovolaemic shock in the trauma patient [Internet]. New South Wales: New South Wales Institute of Trauma and Injury Management; 2007.