Systemic hemostatic agents initiated in trauma patients in the pre-hospital setting: a systematic review

Purpose The effect of systemic hemostatic agents initiated during pre-hospital care of severely injured patients with ongoing bleeding or traumatic brain injury (TBI) remains controversial. A systematic review and meta-analysis was therefore conducted to assess the effectiveness and safety of systemic hemostatic agents as an adjunctive therapy in people with major trauma and hemorrhage or TBI in the context of developing the Italian National Institute of Health guidelines on major trauma integrated management. Methods PubMed, Embase, and Cochrane Library databases were searched up to October 2021 for studies that investigated pre-hospital initiated treatment with systemic hemostatic agents. The certainty of evidence was evaluated with the Grading of Recommendations Assessment, Development, and Evaluation approach, and the quality of each study was determined with the Cochrane risk-of-bias tool. The primary outcome was overall mortality, and secondary outcomes included cause-specific mortality, health-related quality of life, any adverse effects and blood product use, hemorrhage expansion, and patient-reported outcomes. Results Five trials of tranexamic acid (TXA) met the inclusion criteria for this meta-analysis. With a high certainty of evidence, when compared to placebo TXA reduced mortality at 24 h (relative risk = 0.83, 95% confidence interval = 0.73–0.94) and at 1 month among trauma patients (0.91, 0.85–0.97). These results depend on the subgroup of patients with significant hemorrhage because in the subgroup of TBI there are no difference between TXA and placebo. TXA also reduced bleeding death and multiple organ failure whereas no difference in health-related quality of life. Conclusion Balancing benefits and harms, TXA initiated in the pre-hospital setting can be used for patients experiencing major trauma with significant hemorrhage since it reduces the risk of mortality at 24 h and one month with no difference in terms of adverse effects when compared to placebo. Considering the subgroup of severe TBI, no difference in mortality rate was found at 24 h and one month. These results highlight the need to conduct future studies to investigate the role of other systemic hemostatic agents in the pre-hospital settings.


Introduction
Severe trauma is a major global public health issue [1]. According to the World Health Organization, the three most common causes of injury and violence-related deaths specifically road traffic accidents, suicides, and homicides. Furthermore, road traffic crashes and falls are the main causes of traumatic brain injury (TBI) [2]; intracranial bleeding is a possible complication of TBI [3], whose frequency depends on the injury severity [4]. Severely injured trauma patients are characterized by coagulation abnormalities and Annalisa Biffi, Gloria Porcu and Greta Castellini contributed equally to this work as co-first authors.
Silvia Gianola and Osvaldo Chiara contributed equally to this work as co-last authors.
Membership of the Italian National Institute of Health guideline working group is listed in the Acknowledgments.
* Annalisa Biffi annalisa.biffi@unimib.it Extended author information available on the last page of the article a substantially increased mortality rate [5]. Approximately 33% patients with TBI present with coagulopathy, which can increase the hemorrhage expansion and risk of death [4]. Hemorrhagic death generally occurs within the first few hours of injury (median time to hemorrhagic death is 2-2.6 h [6]), while late mortality is defined as death due to multiple organ failure or infection [7]. Over the last decade, increased awareness and improved pre-hospital care may have triggered a reduction in the number of severely injured patients with ongoing bleeding being transferred to a trauma center [8,9]. Early detection and management of traumatic hemorrhage and TBI with intracranial hematoma can improve clinical outcomes [10], and, as suggested by the European guidelines, "current guidelines are needed for the implementation of evidence-based local protocols and algorithms together with parameters to assess key measures of bleeding control and outcome" [1,9]. In trauma patients, interventions for bleeding may be required during the (1) acute phase (0-6 h from injury), with blood loss and shock; (2) intermediate phase (6-24 h from injury), with an increased risk of death due to severe TBI and concomitant physiologic impairment; and the (3) late phase, with the occurrence of inflammatory damages and/or complications [7]. Temporary measures to control hemorrhage in pre-hospital settings may be applied, such as the use of tourniquets, hemostatic dressings and pelvic binders, until definitive care is available. In non-compressible hemorrhage recent evidence suggests the resuscitative endovascular balloon occlusion of the aorta that however their use is still anedoctal [11]. Moreover, the use of pro-hemostatic agents has been suggested as an adjunctive measure to reduce bleeding and prevent trauma-induced coagulopathy [12] and to decrease the size of intracranial hematoma [2]. International guidelines strongly recommend the use of the synthetic lysine analog-tranexamic acid (TXA) in the early care of bleeding trauma patients at risk of significant hemorrhage [1,5]. Thus, off-label early administration of this anti-fibrinolytic agent [1] has become a significant component of major hemorrhage protocols [12], but its infusion after 3 h from injury has been associated with the potential occurrence of adverse effects such as nausea, diarrhea, and stomach ache [1,13]. Ideally, initial administration of TXA should be considered in the pre-hospital phase where possible, as evidenced by the decreasing survival benefit over time [14]. Current evidence suggests that early TXA hemostatic administration may improve outcomes in trauma patients both with hemorrhagic shock and intracranial bleeding [15]. Moreover, in cases of uncontrollable hemorrhage and coagulopathy, the off-label use of TXA has been suggested [1,16]. While recombinant activated coagulation factor VII (rFVIIa) has been rarely investigated and used in the pre-hospital arena [17], TXA was tested by several published studies [18]. This systematic review aimed to assess the effectiveness and safety of systemic hemostatic agents as adjunctive treatment measures initiated in the pre-hospital setting and then continued in the in-hospital setting in patients with major trauma and hemorrhage or with TBI.

Methods
A systematic review was conducted to support the major trauma integrated management guideline panel of the Italian National Institute of Health (Istituto Superiore di Sanità, Sistema Nazionale Linee Guida) in formulating recommendations. Specifically, following the Grading of Recommendations Assessment, Development and Evaluation (GRADE)-ADOLOPMENT approach for guideline creation [19] adopted by the Istituto Superiore di Sanità methodological manual, the panel members decided to follow a structured and systematic adaptation and updating process of the recommended use of systemic hemostatic agents initiated in the pre-hospital setting from the National Institute for Health and Clinical Excellence (NICE) on major trauma (clinical guideline, NG39) [20]. The clinical question addressed in this systematic review was, "Is the use of systemic hemostatic agents clinically effective in improving outcomes in patients with confirmed or suspected hemorrhage in major trauma or with acute TBI?".

Study design
The systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline (Supplemental Material 1) [21]. The study protocol is available at the following link: https:// osf. io/ cmdqk/.

Search strategy
Two professional librarians searched the Embase, PubMed (Medline), and the Cochrane Library databases for randomized controlled trials (RCTs) published up to October 12, 2021, that were related to the use of systemic hemostatic agents introduced in the pre-hospital environment among patients with confirmed or suspected hemorrhage or with acute TBI. Following the GRADE-ADOLOPMENT development process [19], the search strategy included in the high-quality NICE guideline on major trauma from 2015 [20] has been updated.
Full details of the search strategy are reported in Supplemental Material 2.

Study selection
Two independent authors (A.B. and G.P.) completed reference screening and study selection using the following predefined inclusion criteria: (1) population: children, young people, or adults who have experienced a suspected hemorrhage or TBI following a traumatic incident; (2) intervention: administration of a systemic hemostatic agent, (e.g., rFVIIa, TXA, fibrinogen concentrate, prothrombin complex concentrate, or another anti-fibrinolytic agent); (3) comparison: no treatment, placebo, standard care (e.g., crystalloids, blood components), or any other hemostatic agent; and (iv) setting: systemic hemostatic treatment initiated in the prehospital setting. Only RCTs were considered eligible for inclusion; case reports, editorials, and letters were excluded from the search.
When multiple publications for the same trial reported different outcomes or follow-up data, they were counted as a single publication. Discrepancies between reviewers were resolved by consulting a third author (O.C.).

Types of outcome measures and follow-up assessment
The primary outcome was the overall mortality (at 24 or 48 h or 1 month), while the secondary outcomes were causespecific mortality at 1 month (e.g., multiple organ failure [MOF], head injury, hemorrhage), health-related quality of life; any adverse effect (e.g., MOF); blood product use; mortality at 12 months, hemorrhage expansion; and patientreported outcomes.

Data extraction
Two independent authors (A.B. and G.P.) extracted data on publication year, country, and characteristics of the study population (e.g., age, Injury Severity Score [ISS], systolic blood pressure, heart rate, Glasgow Coma Scale, type of injury [blunt %], type of systemic hemostatic agents, and outcome data) using a standardized data-collection form in a spreadsheet format (Microsoft Excel, Redmond, WA, USA). Authors were contacted if the reported data were insufficient or unclear.

Statistical analysis
General characteristics were descriptively synthesized. When sufficient outcome data were available, cumulative analyses were performed. Treatment effects provided an estimate of the risk ratio (RR) for dichotomized outcomes and a mean difference (MD) or standardized mean difference (SMD) value when different outcome measurements were present for continuous outcomes, along with 95% confidence intervals (CI). Heterogeneity between study-specific estimates was tested using chi-squared statistics and measured with the I 2 index and the Cochran's Q test [22,23]. A pooled estimate was obtained by fitting the DerSimonian and Laird random-effects model [24] when several studies were combined; conversely, a fixed-effects model was applied. Subgroup analyses were planned for (1) type of population (e.g., subjects with significant hemorrhage and subjects with TBI); (2) type of injury (e.g., blunt trauma, penetrating trauma); and (3) doses of systemic hemostatic agent when available. All tests were considered significant statistically for p < 0.05. For the primary outcome (i.e., mortality) the clinical relevance was assessed with the minimal important differences (MIDs) taken as risk ratios (RRs) of 0.75 and 1.25. For instance, the RR of 0.75 of mortality is taken as the line denoting the boundary between no clinically important effect and a clinically significant benefit, whilst the RR of 1.25 is taken as the line denoting the boundary between no clinically important effect and a clinically significant harm [20]. The analyses were performed using Review Manager Version 5.4 (Cochrane Collaboration, London, UK).

Risk-of-bias assessment (internal validity)
The Cochrane risk-of-bias tool was used to evaluate the internal validity of the included RCTs. The quality of the evidence was assessed using the GRADE methodology [25] (Supplemental Material 3).
No eligible study was found on the use of fibrinogen concentrates, recombinant activated coagulation factor VII, prothrombin complex concentrates, or other anti-fibrinolytic agents in the pre-hospital setting.

General characteristics
Overall, five RCTs assessed the comparison of TXA and placebo (Supplemental Material 2). The median ISS across studies ranged from a minimum of 12 (interquartile range [IQR] = 5-22) to a maximum of 38.7 (IQR = 25-52.4). Blunt trauma was the most representative feature of patients across studies. In particular, TXA versus placebo was investigated in patients with significant hemorrhage [5,31,36,40] or who were at risk of hemorrhage [30], in patients with isolated TBI [2,26,41], and in those with brain injury and significant hemorrhage [4,34]. The general and clinical characteristics of patients are presented in Table 1.

Overall mortality
All included RCTs (n = 5) reported the overall mortality data. The CRASH-2 and STAAMP trials showed mortality at 24 h, while five RCTs (CRASH-2, CRASH-3, STAAMP, TAMPITI, and the TXA trial) evaluated overall mortality at 1 month.

TXA versus placebo
A statistically significant difference including clinical relevance between TXA and placebo for overall mortality at 24 h was observed (RR = 0.83, 95% CI = 0.74-0.95; two studies, 21,030 patients; Fig. 2A), whereas a statistically but not clinically significant reduction was found at 1 month in the TXA group (RR = 0.93, 95% CI = 0.88-0.97; five studies, 34,873 patients, Fig. 2B). The absolute effects with TXA shown 8 and 12 fewer deaths per 1.000 at 24 h and 1 month, respectively.
In the subgroup analysis for the type of population ( Fig. S1 in Supplemental Material 4), TXA did not reduce the overall mortality at 1 month in patients with

Secondary outcome
Descriptive tables and quantitative analyses for the following outcomes are shown in Supplemental Material 4.
• Cause-specific mortality: Bleeding death at 1 month was significantly reduced in patients who received TXA com- pared to placebo, whereas no differences between the groups at 1 month were detected for MOF, head injury, pulmonary embolism, and sepsis (  (Table S4 and Figs. S7-S16 in Supplemental Material 4). • Blood product use: There was no difference in blood, platelet, or plasma transfusions between groups (Table S5 in Supplemental Material 4). • Other secondary outcomes: None of the RCTs reported the mortality at 12 months or patient-reported outcomes, and no significant reduction was detected regarding hemorrhage expansion (Table S6 in Supplemental Material 4).
The above results suggest a higher safety profile of TXA therapy without a significant difference in complications between the two groups.

Internal validity and certainty of evidence
Five RCTs (CRASH-2, CRASH-3, STAAMP, TAMPITI, and TXA trial) were judged to have good methodological quality (Supplemental Material 3). The certainty of evidence across the RCTs ranged from very low to high quality ( Table 2).

Discussion
To our knowledge, this is the first comprehensive systematic review with a meta-analysis assessing the certainty of evidence by the GRADE approach to prove the efficacy of systemic hemostatic agents as an adjunctive treatment initiated in the pre-hospital setting among traumatic patients with (1) hemorrhage, (2) TBI, or (3) hemorrhage and TBI. With a high certainty of evidence, TXA is associated with a lower risk of mortality of 17% at 24 h (RR = 0.83, 95% CI = 0.74-0.95) and 7% at one month (RR = 0.93, 95% CI 0.88-0.97) when compared to placebo. Although the clinical relevance could be questionable, the anticipated absolute effects at 24 h and 1 month anyway shown a reduction of mortality of 8 and 12 averted deaths per 1000 which may be relevant in a life threatening conditions (i.e., major trauma). Considering the subgroup of severe TBI at 24 h and one month, both failed to find any difference between groups.
Difference between patients with severe hemorrhage and TBI may be explained by the timing of anti-fibrinolytic administration and the severity of TBI [2,26]. TBI, commonly characterized by intracranial hemorrhage, is particularly concerning in this context because it can progress or gradually worsen after hospitalization [43]. One-third of patients affected by severe TBI may develop coagulopathy because of the release of brain phospholipids and tissue factors [44]. However, a recent meta-analysis did not show a strong prevalence of coagulopathy in TBI patients compared to those with injuries in other areas of the body or multiple injuries with TBI [45]. Especially in (20 fewer to 5 fewer) *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) CI: confidence interval; MD: mean difference; RR: risk ratio GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect the early stages [46], TXA administration tends to reduce hemorrhage expansion and mortality caused by bleeding in TBI patients [34]; nevertheless, those affected by intracranial bleeding and neuropathological abnormalities did not show significant benefits of anti-fibrinolytic treatment [26]. Within 24 h of injury deaths are more likely to occur because of excessive bleeding [2]. We can use 24-h mortality as a proxy of the overall mortality effect [6], since the ideal endpoint in literature is considered as mortality within the first 6 h after injury [14,47], in fact the CRASH-2, which represents the most contribution (96.5% of weight of meta-analysis), administered the TXA within 3 h from injury. The effect of systemic hemostatic agents depends on the time between injury and the onset of treatment, which suggests there is a protective effect of these medications on early deaths related to hemorrhage [48].
The treatment benefit is time-dependent, the benefits of TXA therapy seen in the first 3 h may be explained by the increasing levels of plasminogen activator inhibitor-1 and a reduction in fibrinolysis. In CRASH-2, when TXA was administered within 3 h of injury, the risk of death due to bleeding was reduced [29]. However, in the late phase of TXA administration, adverse effects, especially thrombotic disseminated intravascular coagulation, was likely to increase, along with uncontrolled bleeding [29] and a higher risk of mortality [49].
Therefore, its administration may be contraindicated [36]. However, we found no differences between TXA and placebo in terms of adverse effects. In addition, many guidelines have focused on TXA administration in severely injured or bleeding patients [20]. The British Committee for Standards in Hematology recommends the administration of TXA in adult trauma patients with, or who are at risk of, bleeding as early as possible after injury [50], and the STOP the Bleeding Campaign in Europe also recommends its use in bleeding trauma patients within 3 h after an injury and those "en route" to the hospital [51]. On the basis of existing evidence, TXA was also included on the World Health Organization's list of essential medicine for the reduction of death among adult patients with trauma and a significant risk of ongoing hemorrhage [52].
Our results are in agreement with those in the current scientific literature. Specifically, meta-analyses and systematic reviews conducted in the pre-hospital or the hospital setting found that TXA reduces the risk of death from all causes by about 20% [44,53,54] and less hemorrhagic expansion [44] among TBI patients [32,46]; yet some results were discrepant with respect to functional status [32,44,54,55].
No evidence was detected for other adverse events [32,44,46,55], although high-quality RCTs reported reduced vascular occlusive events [44]. However, confounders may have influenced the results, such as blood transfusions realized prior to TXA treatment, although this mechanism is still unknown [53].
This study has some limitations. First, survival bias may be present when patients die before the administration of treatment; thus, the results might not be generalizable across the trauma population [56]. Second, cause-specific mortality can be affected by a subjective and misclassified evaluation (e.g., an in-hospital death for hemorrhage may be caused by hemorrhage and severe TBI [6]). Third, clinical conditions (e.g., cardiac arrest) may have strongly influenced the treatment indication. Statistical analyses may not have taken into account the time-varying treatment effect by assuming uniform effects over time, nor adequately considered risk factors [56], or the study outcomes (e.g., reduction of mortality) [39]. Thus, appropriate strategies may reduce these type of biases [56]. Some studies also did not specify the type of trauma (blunt or penetrating) and did not categorize patients based on their trauma severity, while others performed secondary analyses or re-analyses (e.g., restriction to TBI patients). These factors may have limited the subgroup analysis. Fourth, inclusion criteria were narrow, aiming to balance pros and cons. We considered eligible all RCTs that investigated several interventions delivered in any setting (civilian and military). Thus, although we were able to offer the highest level of evidence (RCTs with high certainty of evidence) achieving more consistency in the interpretation of findings, this approach limited the results as we did not find studies in military setting or studies involving fibrinogen concentrates, prothrombin complex concentrate, or other anti-fibrinolytic agents. The inclusion observational evidence (e.g., non-randomized interventional studies) a different pragmatic approach have been found. Early patient randomization may not have prevented the performance of stratified analysis on the anatomical location of bleeding or other injury [36]. Moreover, difficult diagnoses of traumatic hemorrhages might have reduced the statistical power analysis focused on the anti-fibrinolytic effect on bleeding mortality or the need for blood transfusions [5]. Fifth, heterogeneity was found in reporting the average time from injury to anti-fibrinolytic administration, as well as the time of data collection regarding outcome. Last, the complexity of the trauma cohorts and different choices made by trauma teams in their management across various countries may have affected the overall analyses [30].

Conclusion
Balancing benefits and harms, tranexamic acid initiated in the pre-hospital setting can be used for patients experiencing major trauma with significant hemorrhage. Considering the ideal endpoint in literature is the administration of TXA within 6 h from injury, TXA statistically and clinically reduces mortality rate in trauma patients at 24 h. This result is still statistically significant at 1 month. These results are mainly dragged by the subgroup of patients with significant hemorrhage because in TBI patients subgroup there are no difference between TXA and control. Little-to-no difference in terms of adverse effects was found when comparing TXA and placebo. Further research is needed to investigate the role of other systemic hemostatic agents in the pre-hospital settings. Supplemental