Coagulopathy and damage control resuscitation during transfer due to massive hemorrhage –a single-center retrospective study–

Background: To assess medical procedures, particularly damage control resuscitation, at the time of transfer between hospitals for the purpose of treating massive hemorrhage. Methods: This study used a single-center retrospective observational design, enrolling patients referred to Teine Keijinkai Hospital from another hospital between April 2012 and March 2019 for the treatment of massive hemorrhage. We excluded patients who entered cardiac arrest before arriving at our center. Qualitative or categorical variables were compared using the χ 2 or Fisher’s exact test, as appropriate. Quantitative continuous variables were compared using Mann-Whitney nonparametric tests, as appropriate. Risk factors associated with coagulopathy from univariate analyses (brinogen level £150 mg/dl; prothrombin time-international normalized ratio ³1.5) were entered into stepwise logistic regression analysis. Signicance was dened for values of p < 0.05. Results: Multiple logistic regression analysis revealed trauma (odds ratio (OR) 4.800; 95% condence interval (CI) 2.016–11.433; p < 0.001) and volume of crystalloid solution (OR 1.001; 95%CI 1.000–1.001; p = 0.008) as independent factors associated with coagulopathy. Patients with coagulopathy showed higher 24-h mortality rates (10.9%) than patients without coagulopathy (1.2%; p = 0.021), and cause of death was hemorrhagic shock for all cases of death within 24 h. Conclusion: In our area, withholding intravenous uid to achieve permissive hypotension, early administration of fresh frozen plasma, and use of brinogen concentrate may improve the prognosis of patients with massive hemorrhage undergoing transfer between hospitals. lactate, total bilirubin, creatinine, hemoglobin, platelet, prothrombin time-international normalized ratio partial thromboplastin time, brinogen), Continuous variables are presented as median [interquartile range]. Categorical variables are presented as n (%). Continuous variables are presented as median [interquartile range]. Categorical variables are presented as n (%). Continuous variables are presented as median [interquartile range]. Categorical variables are presented as n (%).

treatment. The time from the scene to a hospital that can provide de nitive care can thus be quite long. Even if the patient is hospitalized, transfer to another hospital may be needed if the bleeding is beyond the capability of the receiving hospital to treat. Because the time from onset to de nitive hemostasis is increased, DCR is likely to become even more important for patients transferred between hospitals.
This study was undertaken to assess medical procedures, particularly DCR, at the time of transfer between hospitals, for the purpose of treating massive hemorrhage in our medical administration area.

Methods
This study applied a single-center, retrospective, observational design. Patients referred to Teine Keijinkai Hospital from another hospital between April 2012 and March 2019 for the treatment of massive hemorrhage were enrolled. We excluded patients who had experienced cardiac arrest before arrival at our center. Patients with "bleeding", "trauma", or "shock" were extracted from the medical records, and patients with causes other than massive hemorrhage were excluded. Massive hemorrhage was de ned as follows: 1) systolic blood pressure ≤90 mmHg at the referring hospital [6,7]; 2) shock index (heart rate / systolic blood pressure) ≥1 at the referring hospital [8]; 3) in cases of maternal hemorrhage, cumulative blood loss ≥1,000 ml or blood loss accompanied by signs or symptoms of hypovolemia [9]; and 4) the judgement of the doctor from the referring hospital (for example, when the amount of bleeding was unknown, measurement of blood pressure was di cult, and a description of massive hemorrhage or shock was present in the medical information).
Information on age, sex, diagnosis, past medical history (liver cirrhosis, blood disease, malignant tumor), oral medications (vitamin K antagonist, direct oral anticoagulants, and antiplatelet agents), vital signs in the referring hospital (Glasgow Coma Scale, systolic blood pressure, heart rate), tracheal intubation at the time of transfer, volume of crystalloid solution before arrival at our center, blood product administration before arrival at our center and dosage units when administered, Abbreviated Injury Scale in cases of trauma, catecholamine administration, tranexamic acid administration, time from onset to arrival at our center, vital signs on arrival at our center (Glasgow Coma Scale, systolic blood pressure, heart rate, body temperature), laboratory test data on arrival (pH, lactate, total bilirubin, creatinine, hemoglobin, platelet, prothrombin timeinternational normalized ratio (PT-INR), activated partial thromboplastin time, brinogen), treatment (surgical exploration, angiography with embolization, endoscopic intervention), amount of blood transfused within 24 h of arrival, outcomes (24-h mortality, 30-day mortality), and adverse events of blood transfusion (transfusion-related acute lung injury, allergic reaction or transfusion related reaction) were obtained.
Qualitative and categorical variables were compared using the χ 2 or Fisher's exact test, as appropriate.
Quantitative continuous variables were compared using Mann-Whitney nonparametric tests where appropriate. Risk factors associated with coagulopathy according to univariate analyses were entered into stepwise logistic regression analysis, and results were presented as odds ratios (ORs) and 95% con dence intervals (CIs). Values of p < 0.05 were considered signi cant. SPSS Statistics version 25 software (SPSS, Chicago, IL) was used for all statistical analyses. Ethics approval was obtained from Teine Keijinkai Hospital, Hokkaido, Japan.

Results
Over the 7-year study period, a total of 1076 patients were transferred to our hospital for the treatment of bleeding, trauma, or shock ( Fig. 1). Among these, 172 patients had shock or massive hemorrhage. After excluding 40 shock patients with other causes of shock, 132 patients with massive hemorrhage were included. Table 1 shows the classi cation of the 132 cases, and Table 2 shows background characteristics of patients.
Before arrival at our center, crystalloid solution had been administered to all cases, and median volume of infusion was particularly high with trauma (1,550 ml; interquartile range (IQR) 1,000-2,250 ml) and maternal hemorrhage (2,500 ml; IQR 1,000-3,000 ml Tranexamic acid administration was performed in 9.8% of all cases and in 13.5% of trauma cases. Mean systolic blood pressure at our center was 100 mmHg (IQR 80-117 mmHg) overall and 99 mmHg (IQR 84-120 mmHg) for trauma.   Table 3 shows the results of blood examination on arrival at our center. Median hemoglobin concentration was 7.5 g/dl (IQR 5.7-9.1 g/dl) in the overall study cohort, which was particularly low for maternal hemorrhage (5.9 g/dl; IQR 5.4-7.8 g/dl) and gastrointestinal hemorrhage (6.8 g/dl; IQR 5.5-8.9 g/dl   Table 4 shows treatments and outcomes. The 24-h mortality rate was 4.5% in the overall study cohort, and all patients who died within 24 h died due to hemorrhagic shock. The 30-day mortality rate was 9.1%. Deaths among patients who survived > 24 h but died within 30 days were attributed to complications such as endstage pancreatic cancer, myocardial infarction, acute kidney injury, or chronic disease such as exacerbation of interstitial pneumonia or chronic heart failure.  Patients with coagulopathy displayed a signi cantly higher 24-h mortality rate than patients without coagulopathy, and cause of death was hemorrhagic shock in all cases of death within 24 h (Table 6).
Considering the cause of hemorrhage, coagulopathy due to trauma tended to be associated with 24-h mortality, but no signi cant difference was observed (21.7% in the coagulopathy group; 0% in the no coagulopathy group, p = 0.077). No signi cant difference in 30-day mortality was identi ed. Patients with coagulopathy received more blood transfusions within 24 h. No differences were noted in rates of transfusion-related acute lung injury, allergic reaction or transfusion-related reaction.

Discussion
According to this survey, many patients with massive hemorrhage who were transferred to our center had coagulopathy, especially due to trauma and maternal hemorrhage. Trauma and volume of crystalloid solution were factors independently associated with coagulopathy. Patients with coagulopathy had higher 24-h mortality, and required more blood transfusions within 24 h.
In trauma patients without brain injury, permissive hypotension to achieve a target systolic blood pressure of 80-90 mmHg is recommended until major bleeding has been stopped in the initial phase following trauma [4,12]. In prehospital settings, intravenous uid administration is recommended to be titrated for a palpable radial pulse using small boluses of uid (250 ml) rather than xed volumes or continuous administration [13]. In this survey, median systolic blood pressure at the time of arrival to our center was 100 mmHg (IQR 80-117 mmHg) overall and 99 mmHg (IQR 84-120 mmHg) for trauma, all of which were within standard values, but were high from the perspective of permissive hypotension. Both trauma and maternal hemorrhage cases received large volumes of crystalloid infusion before reaching our center and many showed coagulopathy at the time of visit, so there was considered to be some room to limit infusion volumes before arrival at our center.
Regarding prehospital administration of blood products, both RBCs and FFP have been reported to show improvements in mortality and coagulopathy among trauma cases [5][6][7]14]. This study targeted transfer cases from hospitals that can administer blood products, not direct transport from the eld. However, the rate of blood product administration was 23.5% for RBCs and 7.6% for FFP in all cases, and 32.4% for RBCs and 8.1% for FFP in trauma cases. In Europe, viscoelastic methods (VEMs), like thromboelastography and rotational thromboelastometry, are recommended for the diagnosis of trauma-induced coagulopathy because of the ability to provide rapid assessment of hemostasis to support clinical decision-making [4]. In Japan, VEM is usually restricted to experimental and research settings in academic hospitals, and VEM equipment is rarely present in most primary-care hospitals. Coagulopathy due to trauma occurs early after trauma [15,16], and in severe trauma such as multiple injuries, the frequency of established coagulopathy on emergency room admission is high [11,17]. In patients with massive hemorrhage, early administration of FFP is rational to prevent coagulopathy. In fact, early administration [18] and high-ratio administration [19,20] of FFP are recommended. In our area, even in environments in which testing for coagulopathy is not possible, blood products including FFP should be administered more often at the time of transfer to a hospital.
On the other hand, patients experiencing massive hemorrhage are under severe time constraints, because transfer to a hospital that can achieve de nitive hemostasis is required as soon as possible. One of the reasons why the rate of FFP administration was low may be that a longer time is needed to thaw the product.
In that respect, brinogen concentrate does not require thawing or cross-matching and allows rapid administration [21], and is useful in prehospital settings [22]. Furthermore, because brinogen levels decrease earlier than any other hemostatic factors in the case of massive hemorrhage [23], supplementation of brinogen is required from early in trauma. Conversely, FFP requires high volumes to maintain brinogen levels [24], so brinogen concentrate can be supplemented with brinogen even in small amounts [21]. This is also advantageous from the perspective of restricted volume replacement for permissive hypotension and prevention of hypocalcemia resulting from the citrate chelation of serum Ca 2+ . Fibrinogen concentrate reportedly carries lower risks of massive transfusion or multiple organ failure than FFP [25]. Fibrinogen concentrate has also been reported to reduce blood loss and total amount of FFP when treating coagulopathy from postpartum hemorrhage [26]. At present, although brinogen concentrate is only approved for bleeding episodes in patients with congenital brinogen de ciency in Japan, due to the abovementioned advantages, some facilities use brinogen concentrate for trauma and obstetric bleeding with the approval of the hospital ethics committee. If administration of brinogen concentrate for patients with bleeding becomes approved in Japan as in many European countries, early correction of coagulopathy should be possible.
Tranexamic acid should be given to bleeding trauma patients as early as possible [27,28], but administration at the referring hospital was limited to 13.5% in our region. In the obstetric setting, tranexamic acid has been shown to be effective, particularly when given early after bleeding onset [29], but administration was limited to 6.1%. On the other hand, routine use of tranexamic acid is not recommended for upper gastrointestinal bleeding [30], and tranexamic acid was used in 14.6% of gastrointestinal bleeding cases. Early treatment with tranexamic acid for trauma and maternal hemorrhage should thus be promoted in our region.
The mechanisms of coagulopathy following trauma are considered to involve tissue hypoperfusion and hypoxia, which in turn induce endothelial damage and activation [31][32][33], and an iatrogenic factor that occurs secondary to uncritical volume therapy leading to acidosis, hypothermia, and hemodilution [33]. Maternal hemorrhage is also associated with a high risk of early coagulopathy, because loss of clotting factors by placental separation and atonic bleeding causes early progression, so dilute coagulopathy is more likely to occur with a small amount of bleeding compared to intraoperative bleeding in other diseases, and obstetric disseminated intravascular coagulation with premature separation of the placenta and amniotic uid embolism shows a very high bleeding tendency [34]. Both trauma and maternal hemorrhage are prone to a high degree of coagulopathy, since the properties cause coagulopathy as well as consequences of bleeding and dilution. In the present study, many coagulopathies were observed due to trauma and maternal hemorrhage.
On the other hand, coagulopathy due to gastrointestinal hemorrhage was not observed. Although many facilities use massive transfusion protocols for early replacement of coagulation factors in cases other than trauma [35,36], the results of this study suggest that administration of a high ratio of FFP to RBCs for gastrointestinal bleeding may not be effective. Many reports have described restriction of blood transfusion as showing better prognosis for gastrointestinal bleeding [37][38][39]. However, since many reports exclude massive bleeding or do not consider severity, transfusion strategies for severe gastrointestinal hemorrhage warrant closer consideration.
This study showed several limitations that merit consideration when interpreting the results. First, this study was a single-center, retrospective study, and the number of subjects was limited, so our results cannot be generalized. However, the results that a large infusion volume of crystalloid solution is associated with coagulopathy and that the presence of early coagulopathy is associated with poor prognosis were the same as reported elsewhere. Few reports have examined whether DCR is performed in prehospital settings, so the result of low compliance with DCR seems to be a problem that is not exclusive to our region. Second, the judgement of the doctor from the referring hospital was adopted in the de nition of massive hemorrhage, because blood pressure or estimated blood loss is di cult to determine due to intra-or retroperitoneal hemorrhage in some cases. Although this involved the inclusion of subjective judgments from doctors at the referring hospital, there was not considered to be any difference in treatment content or prognosis after transfer, because this study did not identify the doctor from the referring hospital or the doctor in charge at that time. Third, this study involved a review of medical records, so if the details of treatment by the doctor from the referring hospital remain unclear, missing values may occur, and results may differ. Fourth, the 30day mortality rate was 9.1% in the overall study cohort and 13.5% in trauma cases. No deaths due to ruptured aortic aneurysm were identi ed. Mortality rates in studies of massive hemorrhage due to trauma reportedly vary from 8.4-37.5% [5,7,12], but were generally higher than the mortality rates in this survey. This was thought to be due to the exclusion of more severe patients, such as those who experienced cardiac arrest before arrival at our center or those in an unstable condition and could not be transferred to the hospital.
Preventing coagulopathy may stabilize the patient into a transferable condition, which does not change the conclusion that early response to coagulopathy is warranted.

Conclusion
Many patients transferred to our center with massive hemorrhage showed coagulopathy at the time of visit. Risk factors for coagulopathy at the time of transfer were trauma and volume of crystalloid solution before arrival. Patients with coagulopathy at the time of visit showed a higher 24-h mortality rate. In our area, withholding intravenous uid for permissive hypotension, early administration of FFP, and use of brinogen concentrate were suggested to improve the prognosis of patients with massive hemorrhage.
List Of Abbreviations DCR, damage control resuscitation; PT-INR, prothrombin time-international normalized ratio; OR, odds ratio; CI, con dence interval; IQR, interquartile range; RBC, red blood cell, FFP, fresh frozen plasma; VEM, viscoelastic method Declarations Ethics approval and consent to participate This study was approved by our institutional review board.

Consent for publication
Not applicable.

Availability of data and materials
The datasets analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
There is no funding source. Figure 1