Geographical Variance in the Use of Tranexamic Acid for Major Trauma Patients

Background and Objectives: The CRASH-2 trial is the largest randomised control trial examining tranexamic acid (TXA) for injured patients. Since its publication, debate has arisen around whether results could be applied to mature trauma systems in developed nations, with global opinion divided. The aim of this study was to determine if, among trauma patients in or at significant risk of major haemorrhages, there is an association of geographic region with the proportion of patients that received tranexamic acid. Materials and Methods: We conducted a systematic review of the literature. Potentially eligible papers were first screened via title and abstract screening. A full copy of the remaining papers was then obtained and screened for final inclusion. The Newcastle–Ottawa Scale for non-randomised control trials was used for quality assessment of the final studies included. A meta-analysis was conducted using a random-effects model, reporting variation in use sub-grouped by geographical location. Results: There were 727 papers identified through database searching and 23 manuscripts met the criteria for final inclusion in this review. There was a statistically significant variation in the use of TXA for included patients. Europe and Oceania had higher usage rates of TXA compared to other continents. Use of TXA in Asia and Africa was significantly less than other continents and varied use was observed in North America. Conclusions: A large geographical variance in the use of TXA for trauma patients in or at significant risk of major haemorrhage currently exists. The populations in Asia and Africa, where the results of CRASH-2 could be most readily generalised to, reported low rates of use. The reason why remains unclear and further research is required to standardise the use of TXA for trauma resuscitation.


Introduction
Trauma is the leading cause of death in developed nations for those under the age of 45 years [1,2]. Among injured patients reaching hospital, haemorrhages are the second leading cause of death-after central nervous system (CNS) injury-and the number one preventable cause of death [3][4][5].
Haemorrhages are complicated in 25-30% of severely injured trauma patients by trauma induced coagulopathy (TIC). TIC is characterised by hypocoagulability and hyperfibrinolysis and carries a poor prognosis and four-fold mortality rate compared to trauma patients without TIC [6]. Tranexamic acid (TXA) is the most widely used antifibrinolytic and it has been hypothesised that its routine use in trauma patients may result in reduced mortality by the correction of hyperfibrinolysis [7].
CRASH-2 is the largest randomised control trial (RCT) examining circulatory resuscitation for trauma patients to date. It enrolled 20,211 adult trauma patients from 274 hospitals in 40 different countries and randomised them into two groups-TXA (10,096) and placebo (10,115). The study found a statistically significant reduction in all-cause mortality in the TXA group (14.5%) compared to the placebo group (16%), with a relative risk (RR) of 0.91 (95% CI 0.85-0.97 p = 0.0035) [8]. CRASH-2 concluded that TXA administration within three hours of injury is safe, cost-effective and may result in a mortality benefit to injured patients [8].
However, the results of CRASH-2 have not been unanimously implemented into practice. Questions have been raised regarding its applicability to developed trauma systems that manage patients with high volumes of plasma and other blood products [9]. Only around half of the participants in CRASH-2 required a transfusion or operation, calling into question subject selection for the study and whether subjects in the study were truly at risk of a significant haemorrhage. A potential complication of TXA use within the trauma setting is thromboembolism [10,11]. Questions have been raised as to whether CRASH-2 adequately assessed for this complication. While the number of clinical events was recorded, no data on the number of patients who were tested for the condition was reported. It has been hypothesized that this is because they were not actively sought after and, thus, may be an underestimation of the true incidence [12]. Further questions regarding limitations of the CRASH-2 trial have since been raised including its approach to randomisation, the comparability of cohorts in regard to both Injury Severity Score (ISS) and shock and mortality rate in patients given TXA after 3 h [13,14].
To assess the translation of such a landmark RCT with positive results into clinical practice, we undertook a systematic review to assess the uptake of TXA into clinical practice after publication of the CRASH-2 trial. Specifically, we aimed to determine the association of regional variation in the use of TXA in trauma patients in or at significant risk of major haemorrhage.

Materials and Methods
All studies describing trauma patients in or at significant risk of major haemorrhage that reported TXA administration since the CRASH-2 trial and described resuscitation with blood and/or blood products and TXA were included. Included studies also had to report the primary outcome of interest, being mortality, at a defined time-point.
Excluded were animal studies, studies that included patients from before the CRASH-2 trial, studies that were not published in English and literature and systematic reviews that did not include original data about the proportion of trauma patients administered TXA.
A senior librarian was consulted to help develop a search strategy. OvidMedline and OvidEmbase were searched using a variety of subject headings and text words to identify relevant papers. Further texts were then identified using a combination of text word searching in Google Scholar and snowballing from reference lists of potentially eligible trials. The final search strategy is included in Appendices A and B.
Once identified, papers were assessed for final inclusion via a multi-tiered approach. Potentially eligible papers were first screened via title and abstract screening. A full copy of the remaining papers was then obtained and read through. Based on full text screening, a final list of papers for potential inclusion was discussed between researchers. Once deemed eligible for final inclusion, the quality of the remaining papers was analysed using the Newcastle-Ottawa Scale (NOS) and reported using their star-based system (Table 1). NOS uses a star-based system to assess the quality of non-randomised control trial studies. NOS assesses three criteria: selection (out of 4 stars), comparability (out of 2 stars) and outcome (out of 3 stars).
Data were initially extracted, and the results were formatted into a table using Microsoft word for Mac version 16.16.8 by a primary reviewer. The results of this table were then discussed among two co-authors. Post-discussion, additional data was searched for and papers were again reviewed by the same primary reviewer. Following this, data was extracted using Covidence Systematic Review Software. Information from the first table and the Covidence Systematic Review Software was then compared and results combined into the two final tables that have been included in this study.

Results
There were 727 papers identified through database searching (Medline, n = 386; and Embase, n = 341) and 23 manuscripts met the criteria for final inclusion in this review. The PRISMA study flow diagram can be seen in Figure 1.
was then compared and results combined into the two final tables that have been included in this study.

Results
There were 727 papers identified through database searching (Medline, n = 386; and Embase, n = 341) and 23 manuscripts met the criteria for final inclusion in this review. The PRISMA study flow diagram can be seen in Figure 1. Out of the 182 papers where the full texts were read, 159 were excluded. 130 of these papers did not meet study design criteria for inclusion in this paper. A further 15 studies were excluded as they focused on a different study population-trauma patients who were not in or at significant risk of a major haemorrhage, or patients who were in or at significant risk of a major haemorrhage but for reasons other than trauma. Nine studies excluded data about patients who were eligible but did not receive TXA. As a result, calculating the proportion of eligible trauma patients who received TXA was impossible, therefore excluding them from this review. Finally, five studies were excluded as they included patients from before the publication of CRASH-2.
The final 23 studies included in the review all reported the proportion of trauma patients in or at significant risk of major haemorrhage who received TXA. A descriptive list of the 23 papers is presented in Table 2. All 23 are retrospective observational studies. Studies based in five different continents were included: Africa (n = 1), Asia (n = 4), Europe (n = 9), North America (n = 8), and Out of the 182 papers where the full texts were read, 159 were excluded. 130 of these papers did not meet study design criteria for inclusion in this paper. A further 15 studies were excluded as they focused on a different study population-trauma patients who were not in or at significant risk of a major haemorrhage, or patients who were in or at significant risk of a major haemorrhage but for reasons other than trauma. Nine studies excluded data about patients who were eligible but did not receive TXA. As a result, calculating the proportion of eligible trauma patients who received TXA was impossible, therefore excluding them from this review. Finally, five studies were excluded as they included patients from before the publication of CRASH-2. The final 23 studies included in the review all reported the proportion of trauma patients in or at significant risk of major haemorrhage who received TXA. A descriptive list of the 23 papers is presented in Table 2. All 23 are retrospective observational studies. Studies based in five different continents were included: Africa (n = 1), Asia (n = 4), Europe (n = 9), North America (n = 8), and Oceania (n = 1). Studies were included from both pre-hospital and emergency departments. Of the final 23 studies, 15 included information on cohorts' average ISS. The majority of these studies had an average ISS of >12, indicating patients had experienced major trauma [15]. Both civilian-and military-based studies were included. The average age of military cohorts was younger in comparison to that of civilian studies, although average ISS was largely comparable. Only five studies reported their facility's definition of 'massive transfusion' and the number of patients who received massive transfusion (MT). However, among studies that included a definition, the definition used was homogenous, with all five studies defining MT as a transfusion of ≥10 units of pRBCs.
Outcomes reported in selected manuscripts are presented in Table 3. The number of patients included in studies showed significant variance ranging from n < 10 to n = 7269. The number of patients eligible who went on to receive TXA also showed significant variance. Uptake was shown to be most predominant in Europe, with studies reporting up to 69% (95% CI: 68-70) of eligible patients receiving TXA [16].
Although included in the final tables, we excluded case studies and case series from meta-analysis and forest plot. We defined case studies and case series as n < 10 and therefore two studies were excluded. These were Aedo-Martin (2016) and Chesters (2015) [17,18].
The NOS chart results are reported in Table 1 below.   Patients for whom the trauma team had been activated with TXA indication, defined as any of the following: (1) Hypotension systolic blood pressure (SBP) <90 mmHg), (2) Receiving blood transfusions, (3) Initiation of the massive transfusion protocol, (4) Actively bleeding or (5) With a clinical concern for ongoing bleeding. If patients arrived >3 h post injury they were deemed ineligible for TXA, however were still included in calculating 'mean age, ISS, MTP activation and blunt vs. penetrating trauma'.      Figure 2 demonstrates the variance in proportion of eligible trauma patients that received TXA. We use authors' self-reported definitions of what constitutes TXA eligibility. There was significant statistical heterogeneity. TXA use was reported to be most predominant in Europe and one study from Oceania. Use in the USA is currently mixed, and Asia and Africa report a low proportion of patients who receive TXA. Overall the reported proportion of TXA use among injured patients at significant risk or suffering major haemorrhage was 41% (95% CI: 25-57).

* 45.5 Undefined
Medicina 2019, 55, 561 13 of 20 Figure 2 demonstrates the variance in proportion of eligible trauma patients that received TXA. We use authors' self-reported definitions of what constitutes TXA eligibility. There was significant statistical heterogeneity. TXA use was reported to be most predominant in Europe and one study from Oceania. Use in the USA is currently mixed, and Asia and Africa report a low proportion of patients who receive TXA. Overall the reported proportion of TXA use among injured patients at significant risk or suffering major haemorrhage was 41% (95% CI: 25-57).

Discussion
This systematic review and meta-analysis of studies reporting on the use of TXA since the publication of CRASH-2 demonstrated substantial uptake of TXA use during trauma resuscitation. However, there is significant global geographical variance in the uptake of TXA for trauma patients. Awareness and understanding of the reasons for this variance are essential for appropriate translations of research findings.
One factor may be the perceived inadequacy of the level of evidence to support the routine use of TXA in the trauma setting. While CRASH-2 is a large multination RCT, and hence level 1 evidence, in a paper published after CRASH-2, 452 trauma surgeons across the USA completed an online survey addressing their use of TXA in the trauma setting. Of the 452, only 38.0% reported that they use TXA routinely in their practice [19]. Of the 72.0% who said they do not routinely use TXA, 47.7% reported that the reason for not using it regularly was 'uncertain clinical benefit'. Furthermore, 10.0% indicated they believe there are better alternatives and 6.1% believe the risks outweigh the perceived benefits [19]. This supports the hypothesis that many physicians in mature trauma systems believe the results of CRASH-2 cannot be extrapolated to their clinical setting.

Discussion
This systematic review and meta-analysis of studies reporting on the use of TXA since the publication of CRASH-2 demonstrated substantial uptake of TXA use during trauma resuscitation. However, there is significant global geographical variance in the uptake of TXA for trauma patients. Awareness and understanding of the reasons for this variance are essential for appropriate translations of research findings.
One factor may be the perceived inadequacy of the level of evidence to support the routine use of TXA in the trauma setting. While CRASH-2 is a large multination RCT, and hence level 1 evidence, in a paper published after CRASH-2, 452 trauma surgeons across the USA completed an online survey addressing their use of TXA in the trauma setting. Of the 452, only 38.0% reported that they use TXA routinely in their practice [19]. Of the 72.0% who said they do not routinely use TXA, 47.7% reported that the reason for not using it regularly was 'uncertain clinical benefit'. Furthermore, 10.0% indicated they believe there are better alternatives and 6.1% believe the risks outweigh the perceived benefits [19].
This supports the hypothesis that many physicians in mature trauma systems believe the results of CRASH-2 cannot be extrapolated to their clinical setting.
Knowledge gaps such as those identified in Napolitano, Cohen, Cotton, Schreiber and Moore [13], in their 2013 paper, are examples of the 'uncertain clinical benefit' to which these US trauma surgeons may have been referring.
Further debate exists around the implication of the results of the CRASH-2 trial. CRASH-2 has led some to the recommendation that TXA should be administered to all trauma patients within 3 h of injury [20]. However, this view is not universal, with even some of the authors of CRASH-2 suggesting that not all trauma patients may benefit from TXA [21].
A large observational study by Valle et al. [22] found increased mortality in trauma patients who received TXA. In their study, they compared 150 TXA patients to propensity-matched equivalents who did not receive TXA. The difference in mortality rates between the two cohorts failed to reach statistical significance. However, in a subgroup of patients who required emergency surgery within 30 min of arrival to hospital, they found a statistically significant two-fold mortality rate in patients receiving TXA compared to their non-TXA equivalents [22]. While this group is not reflective of the average trauma population, their results do add to the argument that TXA may not be universally beneficial for all trauma patients with major haemorrhages.
Another factor contributing to the variance seen may be difficulty in translating the evidence into clinical practice. It is estimated that it takes an average of 17 years for evidence-based research to reach clinical practice [23]. While this is a rough and generalised average, it is reasonable to consider that the evidence from the CRASH-2 trial, published nine years ago in 2010, is in the 'lag' period between publication and clinical implementation. In their study Coats, Fragoso-Iniguez and Roberts [16] reported that TXA use increased every year, from 0% in 2010 to nearly 80% in 2016 [16]. This suggests that the results of CRASH-2 have not yet been fully realised and are still being translated into clinical practice.
This systematic review and meta-analysis have shown that the implementation of TXA since CRASH-2 has been most substantial in the UK. Following publication, UK-based participants of the CRASH-2 trial extensively promoted the trial using social-media, multi-media and medical education websites [24]. In 2011, a short Claymation video called "TRANMAN" was published by the London School of Hygiene and Tropical Medicine, promoting the use of TXA and encouraging people to review the results of CRASH-2 [25]. This was followed up by a song performed by a community choir, in which the lyrics encourage doctors to "save my life, use tranexamic acid" in response to a variety of hypothesised injuries [26]. Finally, a comic strip was commissioned and distributed to all EDs in the UK, in which emergency doctors use TXA for a variety of trauma scenarios, exclaiming its benefits [27]. The increased speed of translation of CRASH-2 into clinical practice in the UK, compared to other countries, may be due to the extensive promotion by some of the UK-based participants.
It must also be noted that the UK implements a method of additional funding via a "Best practice tariff" (BPT). A level 1 BPT of £1406 is made to hospitals for trauma patients with an ISS >8 and a level 2 BPT of £2819 for patients with an ISS >16, given that they fulfill 6 criteria, one of which is that TXA must be administered within 3 h of injury for patients requiring blood products [28,29]. It is reasonable to hypothesise that incentivisation of TXA administration with additional funding via a BPT may be one of the reasons for increased TXA administration within the UK.
The final factor we believe may be contributing to geographical variance is the ability of trauma systems to implement TXA into their clinical practice. While the WHO has added TXA to its list of essential medicines and it is widely available around the world, patients must present to a facility able to transfuse it within three hours of injury [8,30]. Despite being conducted in a developed nation with a mature trauma system, Bardes et al. [31] reported that 30.4% of all patients with an indication for TXA arrived to ED outside the 3 h window, deeming them ineligible for treatment [31]. Injured patients in rural settings and in nations with less mature trauma systems take longer to reach tertiary care facilities equipped to deal with their injuries [31,32]. This may limit the system's ability to implement TXA, despite obvious clinical indications. Multiple nations have begun to implement pre-hospital TXA administration to overcome this barrier, however this practice is not universal and requires an advanced pre-hospital trauma system. This is the first systematic review examining geographical variance in the use of TXA for trauma patients. CRASH-2 was published nine years ago and uptake appears to still be increasing. Therefore, it is plausible this data is an underestimation of current TXA use. It is also plausible that this systematic review suffers from publication bias. While grey literature was searched, only one unpublished paper was included in the final review. Published papers are more likely to show extreme results and may not be a true reflection of current practice. Furthermore, results of publications from single sites cannot be extrapolated and assumed to represent the wider geographical region. We know there is significant variation in the use of TXA based upon geographical location, it may also be possible that there is significant variation within these regions. Thus, results from published studies may only reflect use at those individual sites, rather than the region as a whole. We also acknowledge that the number of manuscripts published may not necessarily reflect drug use within a country.
This systematic review cannot postulate underlying reasons why significant geographical variance exists. Future research is needed to determine the cause of the variance seen in clinical practice. This systematic review also cannot determine the clinical benefit or harm of TXA use for trauma patients. Future high-quality evidence is needed to answer this question.
Two multinational RCTs are currently underway examining the use of TXA for trauma patients. These are the Pre-hospital Anti-fibrinolytics for Traumatic Coagulopathy and Haemorrhage Study (PATCH) and Study of Tranexamic Acid During Air Medical Pre-hospital Transport (STAAMP) trials. The Pre-hospital Anti-fibrinolytics for Traumatic Coagulopathy and Haemorrhage Study (PATCH) is an international, multicenter, double-blinded, placebo-controlled trial examining TXA in advanced trauma systems based out of Australia and New Zealand. PATCH will enroll 1200 severely injured patients deemed to be at risk of acute traumatic coagulopathy (ATC) using the coagulopathy of severe trauma (COAST) scoring system. The primary outcome of the PATCH trial will be the proportion of patients with a favourable outcome at 6 months, as defined by an extended Glasgow outcome score (GOSE) of 5-8 [33].
Study of Tranexamic Acid During Air Medical Pre-hospital Transport (STAAMP) is a US-based multicenter, double-blinded, placebo-controlled trial examining TXA in trauma patients. STAAMP will enroll 1000 severely injured patients at risk of a significant haemorrhage as defined by a pre-hospital systolic blood pressure (SBP) <90 or HR >110. The primary outcome is all-cause mortality at 30 days [34].
Both trials are currently in the recruiting stage but, upon completion, will potentially provide high-quality evidence regarding the benefits and harms of TXA implementation in a modern trauma system. It is hoped that this evidence will provide definitive answers to the debate that currently exists.

Conclusions
A large geographical variance in the use of TXA for trauma patients in or at significant risk of a major haemorrhage currently exists. However, the reason why remains unclear. Further studies are needed to explain the cause of this variance.

Author Contributions:
The study was conceptualized and the methodology was determined by all authors. Formal analyses were undertaken by K.W. and B.M. Writing-review and editing was performed by all authors.
Funding: This research received no external funding.