Prehospital Transfusion Outcomes in Traumatic Hemorrhage: a Prospective Cohort Study

Rationale Trauma is one of the leading causes of death among younger people. According to the World Health Organization (WHO) injuries are responsible for approximately 4,4 million deaths around the world every year. Most preventable deaths following trauma are due to hemorrhage and occur rapidly after injury. Previous research has shown that early initiation of transfusion reduces mortality in hemorrhaging patients. This highlights the need for timely appropriate management of bleeding patients in the pre-hospital setting.

Uncertainty remains regarding the actual benefit of prehospital transfusion. While early prehospital transfusion appears to be a logical step in achieving hemostatic resuscitation at an early stage, the available evidence is inconsistent. Studies report conflicting results and lack robust data demonstrating a clear survival benefit. This uncertainty has limited the development of clear recommendations and guidelines on prehospital transfusion. Additionally, the use of prehospital blood products poses logistical and financial challenges in a setting with already limited resources.

In the Netherlands the Helicopter Emergency Medical Service (HEMS), also called the 'Mobiel Medisch Team' (MMT) is an addition to the usual ambulance care. The Dutch HEMS consists of four bases and is deployed when a patient's health is acutely and severely threatened. The decision to deploy the HEMS can be made by the dispatcher in the emergency control room or by care providers on the scene. Currently, all Dutch HEMS carries 2 units of red blood cells (RBCs), 1g tranexamic acid (TXA) and 2g of fibrinogen. Since this year 2 units of OctaplasLG Lyo freeze dried plasma (FDP) were added as a transfusion product at two HEMS stations. On the HEMS bases equipped with both RBCs and plasma, products are transfused in a 1:1 ratio, in accordance with current literature.

Many hospitals and HEMS services have undertaken studies to evaluate prehospital transfusion programs. However, such studies have not yet been conducted within the Dutch HEMS population, resulting in a significant gap in knowledge regarding optimal pre-hospital transfusion strategies and subsequent in-hospital management of hemorrhaging trauma patients in the Netherlands. Therefore, this study proposes to assess prehospital transfusion practices in the Netherlands to identify the optimal transfusion strategy.

Objectives The primary objective is to evaluate the impact of the differing transfusion strategies employed across Dutch HEMS bases for trauma patients with (signs of) massive hemorrhage on 24 hour mortality.

Secondary Objective(s) are:

1. Identify and describe prehospital and in-hospital characteristics of trauma patients suffering from severe hemorrhage.
2. Identify and describe short- and long-term clinical outcomes.
3. Determine whether specific patient characteristics can predict a greater benefit from prehospital transfusion

Study type Multicenter prospective cohort study

Study population Adult (>16 years) trauma patients with (signs of) massive bleeding who are attended to by the HEMS and brought to a level 1 trauma center and/or received prehospital transfusion products. Patients will be assessed for eligibility to enter the trial according to the criteria set out below. If patients are eligible for inclusion, after initial screening they will be enrolled under exemption of consent. Patients are considered eligible for inclusion if they fulfil all the inclusion criteria and none of the exclusion criteria.

Inclusion criteria

In order to be eligible to participate in this study, a participant must meet all of the following criteria:

1. Adult (age ≥ 16 years)
2. Patients who received prehospital blood products from the HEMS
3. Patient has suffered traumatic injuries
4. Patient is transported to a level 1 trauma center

Exclusion criteria

A potential participant who meets any of the following criteria will be excluded from participation in this study:

1. Physician in prehospital setting decides injuries are incompatible with life before treatment started
2. Transfer from other hospital

Methods Directly after arrival of the HEMS team on scene, the physician will assess and treat the patient according to standard HEMS guidelines and Advanced Trauma Life Support (ATLS) standards. After life threatening airway and breathing problems are ruled out or resolved, a patient will be assessed for traumatic hemorrhage. External bleeding will be controlled by direct compression or tourniquets and a large bore intravenous (IV) access will be obtained. If a patient is suspected of massive hemorrhage transfusion with packed red blood cells (pRBC's) and (if available) FDP will be initiated, according to the HEMS massive transfusion protocol. Additionally, 1 gram of TXA and 2 grams of fibrinogen will be administered. Once a patient has arrived at a level 1 trauma center local protocols will be followed.

If a patient is suspected of massive hemorrhage the HEMS physician will treat the patient according to standard clinical care. After a patient is admitted to the emergency department (ED) and the HEMS physician is no longer involved in patient care, the physician will complete routine medical documentation in the HEMS electronic patient record (EPD). A limited number of study-specific variables related to prehospital care that are not routinely captured in the HEMS EPD will be recorded in castor by the HEMS physician. After inclusion in the castor database the research team will extract all remaining study data on the clinical course and outcomes from the HEMS EPD, the emergency medical service (EMS), the patients hospital medical record and the Dutch National Trauma Registry. Patients will be followed for 30 days or death whichever comes first.

Burden and risks The burden and the risks are very low in this study. Standard of care will not be influenced by this study. No additional procedures will be performed. Patients will not be asked to fill in questionnaires or participate actively in the research. The clinical outcome data will be anonymized and stored on castor. The intended study population are participants who have been affected by severe injuries that result in uncontrolled bleeding. Patients that are eligible for inclusion in this study will be incapacitated at the time of screening for this study for multiple reasons including the circumstances of the traumatic event, the nature of their injuries, having severe pain, being under anesthesia, having reduces consciousness and/or the use of strong analgesia. Given that severe traumatic injuries cannot be predicted or anticipated, there is no possibility to identify individuals before the event. Consequently, screening cannot take place prior to the trauma, which makes it necessary to include subjects who are incapacitated.

Recruitment and consent Due to the emergency nature of this trial participants will be enrolled in the study through an opt-out providing participants with a chance to refuse inclusion in the study at a later time. The intended study population consists of patients who have been affected by severe injuries that result in uncontrolled bleeding. Most patients will be incapacitated at the time of screening for this study for multiple reasons, including the circumstances of the traumatic event, the nature of their injuries, having severe pain, being under anesthesia, having reduced consciousness and/or the use of strong analgesia. Additionally, traumatic injury can't be predicted or foreseen, making it impossible to seek consent in advance. Participants will therefore be incapable of giving informed consent prior to enrollment in the study. In the rare event that a participant retains the capacity to give consent in the prehospital setting, their clinical condition will still require immediate resuscitation. It would be inappropriate to delay life-saving treatment and transport in order to seek informed consent. For the same reason it would be clinically unjustifiable to delay treatment until full informed consent can be obtained by proxy An opt-out procedure was chosen, because requiring informed consent after recovery could introduce substantial selection bias. Patients with the most severe injuries are more likely to remain critically ill for prolonged periods. It is often uncertain whether or when they will regain the capacity to provide informed consent. Furthermore, legal representatives are often confronted with major emotional distress during this period and may not be in a position to adequately consider research participation. Being approached about research during this period may be perceived as an additional burden. As a result, a proxy may be less likely to engage in a consent procedure, regardless of the minimal burden associated with this study. Requiring active informed consent would therefore be expected to result in systematic underrepresentation of the most severely injured patients, particularly those with prolonged critical illness. Therefore, the study population would no longer reflect the population of patients with severe traumatic hemorrhage, possibly compromising the validity, generalizability and clinical relevance of the study findings. To balance scientific validity with respect for a patients autonomy, surviving patients will be informed about study participation as soon as possible and provided with a 30-day opportunity to op-out from participating in the study.

Patients who survive the acute phase will receive study information as soon as reasonably possible after admission and will be offered the opportunity to object to the use of their data within 30 days. The information letter will explain the purpose of the study, the type of data collected and the opt-out procedure. If after 30 days no response is received, the patient will remain included in the study. Data from participants who opt out will be removed from the study database and will not be included in the analyses.

In the event that a participant dies during the study, there may have been no opportunity for contact with the research team. In this instance, we recognize that approaching a proxy to ask their opinion is likely to be distressing and without benefit. It is unethical to burden grieving relatives with a decision that had no impact on already performed treatment and only pertains the use of already collected and coded data. The data will be coded so use of data has no implications for the patient or legal representatives. Therefore, we will not approach a proxy to seek their opinion when a participant dies.

Analysis Data will be analyzed using R software and R-studio (R Foundation for statistical Computing, Vienna, Austria). Normality of continuous data will be visually assessed and tested with the Shapiro Wilk test. A two sided p-value <0.05 will be used as a threshold for statistical significance in all statistical tests.

Variables with more than 10% missing values will be considered for multiple imputation following the predictive mean matching method, using ten imputations. Imputation will only be performed if it can be assumed that values were missing at random.

Patients will be divided into groups related to the transfusion strategy. Continuous data will be reported as means with a standard deviation (SD). For non-parametric continuous data median and inter quartile range (IQR) will be used. For categorical data, numbers and frequencies will be reported per subgroup. Descriptive analyses will be performed for all baseline characteristics and clinical outcome measures. Differences between groups will be assessed with the following analyses: for continuous parametric data a T-test will be used. If continuous data is skewed a Wilcoxon rank sum test will be used. Categorical data will be analyzed with a chi squared or Fishers exact test.

The primary analysis investigating the effect of different transfusion strategies (RBCs + plasma vs RBCs only) on 24-hour all-cause mortality will be conducted using a log-binomial regression model.

24-hour mortality will also be assessed using a multivariable logistic regression model and a Kaplan Meier survival curve. Confounders for all analyses will be selected based on variables with a statistically significant (p < 0.10) or clinically relevant differences in baseline characteristics.

Secondary outcomes will be assessed as well. All-cause prehospital mortality, 1 hour mortality and 6 hour mortality will be assessed with a logistic regression and 30 day mortality will be assessed with a cox proportional hazard model. To assess changes from baseline characteristics, blood gas results and lab values overtime a mixed model for repeated measures will be used. A logistic regression will be performed to assess the amount of patients transfused in a 1:1:1 ratio in the ED, the amount of patients with TIC in the ED and at six hours and the amount of patients transfused with the need for massive transfusion in the ED and at the first 24 hours. Length of stay in the ICU and in-hospital will be assessed using a negative binomial regression analysis.

Privacy protection Research data will be stored under a code number that relates to individual participants. This code consists of the study acronym and four numbers for patient ID. A subject identification code list will be used and saved in a separate file. Only the code number will be used for study documentation, annual progress reports, and research publications. The principal investigator (PI) and research assistant are the only people with the information linking individual participants to study code numbers. This data will be stored during the investigation and will be disposed of 10 years after completion of the study. The pseudonymized database will be stored on castor, for the duration of the study. Only researchers involved in the study will be granted access to the castor. The handling of personal data will be in compliance with the EU General Data Protection Regulation (in Dutch: 'Algemene Verodening Gegevensbescherming (AVG)).

Handling and storage of data Once the data is obtained from the medical records, the principle investigator will remove all identifiable characteristics and every case will be given a subject code, that consists of the study acronym and three numbers for patient ID. Researchers involved in the project will be granted access to the database in castor. In line with Erasmus MC guidelines, data will be kept 10 years after it is collected.