This review details the current knowledge in massive hemorrhage with regard to the pathophysiology and laboratory assessment of the coagulation disturbance, the role of plasma and platelet transfusion, the role of pharmaceutical strategies, and the clinical value of having a massive transfusion protocol. The bulk of the recent medical literature on this topic stems from the study of military and civilian trauma; however, where applicable, reference to postcardiac and post-noncardiac surgery and postpartum hemorrhage will also be discussed.
The coagulation disturbance in massively bleeding patients is more than just the result of consumption of clotting factors at sites of injury and dilution from the infusion of fluids and RBCs. Even before substantial amounts of fluid resuscitation and RBC transfusion, one-quarter of trauma patients already have abnormal coagulation parameters. There is an apparent role for the activation of protein C, hypofibrinogenemia, and fibrin(gen)olysis in the coagulation disturbance after trauma and massive hemorrhage. None of these 3 disturbances would be completely mitigated by the use of plasma alone, suggesting that there may be an opportunity to improve the survival of these patients with alternative strategies such as antifibrinolytics. Currently, we are also unclear as to the most accurate and timely method to assess the coagulation disturbance in the massively bleeding patient.
Despite numerous retrospective cohort studies evaluating 1:1 plasma to RBC formula-driven resuscitation, the overall clinical value of this approach is unclear. Studies have even raised concerns regarding a potential increase in morbidity associated with this approach, particularly for patients over-triaged to 1:1, in whom a massive transfusion is unlikely. We have high-quality data that argue against a role for recombinant factor VIIa that should prompt removal of this strategy from existing massive hemorrhage protocols. In contrast, we have high-level evidence that all bleeding trauma, cardiac surgery, and noncardiac surgery patients should receive tranexamic acid as soon as possible. The role for tranexamic acid in the management of postpartum hemorrhage will be clarified by the multicenter, randomized World Maternal Antifibrinolytic (WOMAN) trial. Recommendations regarding when to administer this therapy must be included in all hemorrhage protocols. If we are to improve the care of massively bleeding patients on a firm scientific ground, we will need large-scale randomized trials to delineate the roles of coagulation replacement and laboratory monitoring. But even before these trials are completed, it is clear that a massive hemorrhage protocol is needed in all hospitals that manage bleeding patients to ensure a prompt and coordinated response to hemorrhage.
Recently, we have seen a flurry of reports on the management of massive hemorrhage, particularly in the setting of trauma, postcardiac surgery, and obstetrical hemorrhage. We have progressed substantially in our understanding of the pathophysiology of the coagulopathy associated with traumatic injury, particularly in regard to an important role of activated protein C, hypofibrinogenemia, and fibrin(gen)olysis. Nonetheless, with the exception of the clear utility of tranexamic acid in bleeding patients, we are no further ahead in our understanding of the role of plasma and coagulation concentrates and laboratory coagulation assessment due to the poor quality of the data available in the literature. However, results from studies evaluating the effectiveness of massive hemorrhage protocols suggest an improvement in outcomes for such patients, possibly due to standardization of care and rapid access to blood.
Pathophysiology of the coagulation disturbance in massive hemorrhage
In the last decade, we have come to realize that the coagulation disturbance in trauma patients is more than just the result of consumption of clotting factors at sites of injury and dilution from the infusion of IV fluids and RBCs. Postcardiac surgery hemorrhage also presents a unique complex bleeding challenge due to multiple defects in hemostasis, including but not limited to preoperative use of anticoagulants and antiplatelet agents, on-pump heparin use, and the activation of inflammatory, hemostatic, and fibrinolytic pathways. We have less knowledge regarding the pathophysiology of the coagulation disturbance for bleeding patients after noncardiac, high-blood-loss surgeries and in the setting of postpartum hemorrhage.
In 2003, Brohi et al analyzed a cohort of 1088 trauma patients and found that 24% had a coagulopathy on arrival.1 Patients presenting with these abnormal laboratory test results had a higher rate of death (46% vs 11%, P < .001). This study was the first to raise suspicion that hemodilution was not the sole cause of the coagulopathy. In addition, those investigators found no association between the volume of fluid administered and the development of abnormal coagulation test results. The incidence of coagulopathy was strongly associated with the injury severity score, suggesting that the degree of tissue injury and/or hypoperfusion may be causative factors in this process.
Several studies have confirmed that severely injured trauma patients have activation of the protein C pathway that results in an “anticoagulant ” effect.2–4 A prospective cohort study of 203 major trauma patients found that patients with tissue hypoperfusion and severe injury had activation of protein C, with subsequent inactivation of factor V and VIII and derepression of fibrinolysis.4 Protein C is activated by thrombin, thrombomodulin, and endothelial protein C receptor. Hypoperfusion is believed to result in increased expression of thrombomodulin on the surface of endothelial cells. The thrombomodulin binds thrombin and the complex activates protein C. Activated protein C proteolytically inactivates factor V and VIII, resulting in impairment of clot generation. In addition, activated protein C depletes plasminogen activator inhibitor (PAI-1). Depletion of PAI-1 results in unrestrained tissue plasminogen activator, resulting in plasmin generation and clot lysis (fibrinolysis). In addition, plasmin is able to degrade fibrinogen (fibrinogenolysis), further depleting the fibrinogen reserves. Fibrinogen has recently been recognized to be a critical factor in the bleeding patient, participating in both primary hemostasis (by bridging between platelets via glycoprotein IIb/IIIa) and secondary hemostasis (fibrin formation). More than 15 years ago, Hiippala et al showed in noncardiac surgery patients that fibrinogen was the first factor to reach a set “critical level” (< 1.0 g/L) at 142% of total blood volume loss compared with platelets decreasing below a critical level (50 × 109/L) at 230% of total blood volume loss.5 Variable incidence of hyperfibrinolysis in trauma patients has been reported. When measured by thromboelastography (TEG) in a cohort of 161 trauma patients, the incidence was only 2.5%, but was associated with a mortality rate of 67%.6 Similarly, in a study of 334 major trauma patients, hyperfibrinolysis was observed in 6.8% of patients, and the mortality rate was 86%.7 The lethality of hyperfibrinolysis emphasizes the importance of early recognition of this entity, early administration of antifibrinolytic drugs, and possibly the use of fibrinogen replacement.8 Most recently, investigators have implicated platelet dysfunction as a previously unrecognized defect in trauma patients9 and an inverse association between baseline platelet count and mortality rate at 24 hours has been found.10
Importance of identifying clinical factors elevating the risk of severe hemorrhage
Whether it is a patient presenting to the operating room electively or arriving in the trauma room, it is extremely important to identify any patient factors that could increase the likelihood of hemorrhage and to modify the transfusion therapy plan accordingly. A history should be sought for congenital (particularly the not uncommon VWD) and acquired bleeding disorders (eg, vitamin K deficiency). In addition, effort should be made to determine whether the patient was on antiplatelet and/or anticoagulant therapies, especially given the increasingly widespread use of novel anticoagulants (eg, rivaroxaban and dabigatran). Atrial fibrillation is more frequent after the age of 45 years, 2% in one series, of whom 30% were on warfarin.11 In the same report, the prevalence of atrial fibrillation was even higher in patients over the age of 75 years (8%). Given that the newer anticoagulants (eg, rivaroxaban and dabigatran) are oral, the prevalence of anticoagulant therapy may increase in patients with atrial fibrillation due to better acceptance by patients and physicians. In addition, the most recent guidelines for the management of atrial fibrillation recommend oral anticoagulants for most patients, including all patients more than 65 years of age or having even one risk factor for stroke (ie, congestive heart failure, hypertension, diabetes, prior stroke/transient ischemic attack).12 Therefore, very few patients with atrial fibrillation will not be on an oral anticoagulant. For unconscious patients, clinicians must have a high index of suspicion for antiplatelet and anticoagulant therapy, particularly if the baseline coagulation testing is unexpectedly abnormal.
In contrast to warfarin, for which prothrombin complex concentrates (PCCs) and IV vitamin K are available for the immediate and rapid reversal of the anticoagulant effect in the setting of massive bleeding, no antidote is available for the new oral anticoagulants. These agents have presented new and challenging problems for trauma physicians.13,14 Fortunately, in the setting of normal renal function, these agents have a short (< 24-48 hours) duration of action. In settings such as intracranial hemorrhage or massive bleeding when immediate reversal is required, activated PCCs, PCCs, and recombinant factor VIIa have been used. There is evidence that activated PCCs and PCCs are effective at reversing the coagulation test result for healthy volunteer subjects on rivaroxaban,15,16 although clinical evidence supporting a clinical hemostatic effect has not been verified. In contrast, these studies provide less support for the use of these agents for the reversal of dabigatran-related hemorrhage, although some evidence for a hemostatic effect was seen in animal models.17 Despite the lack of supporting data, PCCs and recombinant factor VIIa will likely be used for the management of severe bleeding from dabigatran while we await additional clinical research.18 In addition, general hemostatic therapies such as topical hemostatic agents, antifibrinolytic agents, surgical hemostasis, and interventional radiology procedures should not be forgotten in the management of these patients. Increased renal clearance of the drugs should also be encouraged with diuresis and, when required for severe hemorrhage, hemodialysis.19
Role of laboratory assessment of the coagulation cascade in bleeding patients
There are 2 major limitations to standard laboratory measures of the coagulation cascade: (1) the results are often not available when decisions are being made regarding blood component replacement and (2) there is very little evidence to support these metrics to guide clinicians on when (and how much) they should transfuse in terms of plasma and other blood products. Chandler et al have shown that the results from complete blood count, international normalized ratio (INR), and fibrinogen tests can be provided within 15 minutes with minor alternations to laboratory procedures as part of a “rapid hemorrhage panel” addressing the first concern.20 This turnaround time is comparable to rapid TEG and rotational TEG.21 Given these rapid turnaround times, blind administration of blood and blood products should only be necessary in the most extreme bleeding situations.
The second limitation of routine coagulation testing is even more problematic. First, the platelet count provides no information on platelet function and clinicians often blindly assume that all platelets counted are functional. However, these patients are often on antiplatelet medication or have post-bypass platelet dysfunction necessitating a measure of platelet function, not just number. Second, both the INR and the partial thromboplastin time are suboptimal measures of thrombin generation and fibrinogen, and measurement of fibrinogen by the Clauss method is usually the last parameter to be reported out by the laboratory (and probably one of the most important pieces of information when managing a massive hemorrhage). Third, the standard tests of coagulation provide no information regarding fibrinolysis.
These limitations have prompted a resurgence in interested in TEG, whether by TEG or rotational TEG. An excellent update on the use of TEG and its role in guiding transfusion practice has recently been published.22 The use of TEG has several benefits, including: sample run on whole blood (avoiding need for centrifugation and plasma separation), can be run as a near-patient test, provides information on coagulation factors, platelet function, fibrinogen level, and fibrinolysis. In addition, the time to results is very rapid. Published reports are available on the use of TEG to guide the management of trauma patients,23,24 cardiac surgery patients25 and during postpartum bleeding.26 The evidence on the use of TEG for guidance of transfusion therapy will be discussed below.
For the majority of clinicians managing these bleeding patients, the current test list only includes the hemoglobin level, platelet count, INR, partial thromboplastin time, and fibrinogen. There are no high-quality data about where to target these measures during a massive hemorrhage. Numerous guidelines have been published to provide guidance on a reasonable approach as to when RBCs, plasma, platelet, and cryoprecipitate transfusions should be administered to the massively bleeding patient in the absence of high-quality data.27–30
Resuscitation with ratio-based plasma infusion
In 2007 and 2008, Borgman and Spinella published 2 studies detailing the remarkable beneficial effect of high ratios of plasma to RBCs on the mortality of injured patients in Iraq.31,32 They concluded that the use of a ratios of 1:1 plasma to RBCs resulted in an absolute risk reduction in mortality of 55% and would assist with the cessation of hemorrhage and therefore a reduction in the need for RBC transfusion. These 2 studies were retrospective in nature, with expected differences in the baseline characteristics between patients receiving high- and low-ratio plasma therapy. Patients were not managed with immediately available thawed plasma, so both studies were affected by survivorship bias. Because plasma has a processing/transport time of approximately 93 minutes33 until commencement of infusion after arrival to hospital, nonsurvivors are more likely to die without having the opportunity to receive plasma. Snyder et al were the first to provide evidence that these studies likely reflected survivorship bias rather than a true beneficial effect of plasma.33 Rajasekhar et al subsequently published a systematic review on this topic, including 11 studies (3 prospective, 7 retrospective, and 1 case-control)34 and concluded that there was insufficient evidence to support the use of fixed ratios in massively transfused trauma patients. In addition, similar reports on the benefit of higher ratios of platelets to RBC units have been published, with similar limitations.35 Nonrandomized trials will not provide us with the definitive answer to this important question. Two randomized trials are under way to assist with answering this question, the Trauma Lab versus Formula Pilot Trial (TR-FL)36 and the Pragmatic Randomized Optimum Platelet and Plasma Ratios (PROPPR) trial. The PROPPR trial will be the first prospective trial to address the role of early platelet transfusion in the management of massively bleeding trauma patients.
The US Army was the first to change practice and commence managing patients with thawed plasma at arrival to hospital. In 2010, they reported on 777 massively transfused trauma patients cared for before and after implementation of the 1:1 plasma to RBCs protocol.37 Despite clear adoption of the protocol, there was no mortality benefit observed. In addition, despite receiving a median of 6 additional units of plasma and 1 additional apheresis platelet transfusion, the use of RBCs actually increased, providing no evidence that this therapy provided hemostatic benefit.
Researchers have also raised concerns about the collateral damage of 1:1 protocols.38–41 Two issues have been raised: (1) over-triage to 1:1 will result in patients receiving plasma when they need none and (2) for patients who are massively transfused, the extra plasma will result in complications of fluid overload (ie, edema, abdominal compartment syndrome, and respiratory compromise) and will increase the risk of multiple organ failure. A report of 1716 nonmassively transfused trauma patients compared patients managed with 1:1 compared with no plasma.41 Plasma administration was associated with a substantial increase in complications, including a 12-fold increase in adult respiratory distress syndrome, a 6-fold increase in multiple organ dysfunction, a 4-fold increase in pneumonia and sepsis, and no improvement in survival. This was confirmed in a study of 1788 nonmassively transfused trauma patients.40 Finally, it is critically important to cease transfusion of blood components (ie, plasma, platelets, and cryoprecipitate) once the clinical team has achieved hemostasis. At this point in the resuscitation, further blood products likely only present risks (eg, fluid overload and transfusion complications) without clear benefit.
Alternative (and adjunctive) strategies to ratio-based plasma infusions
Numerous alterative or adjunctive strategies have been studied in the massively bleeding patient, including but not limited to antifibrinolytics, resuscitation with coagulation concentrates, and goal-directed individualized care based on laboratory coagulation assessment. The use of recombinant factor VIIa will not be discussed in this review due to the lack of benefit seen in clinical trials (including trauma)42,43 and clear evidence of harm.44,45 Recombinant factor VIIa should be removed as an adjunctive strategy from massive hemorrhage protocols.
We have sound evidence for the use of tranexamic acid for all trauma patients presenting with significant hemorrhage.46 The Clinical Randomisation of an Antifibrinolytic in Significant Haemorrhage (CRASH-2) trial randomized more than 20 000 patients to either tranexamic acid or placebo. Tranexamic acid reduces the risk of death (odds ratio [OR] = 0.91, 95% confidence interval [CI], 0.85-0.97, P = .0035) and death from hemorrhage (OR =0.85, 95% CI, 0.76-0.96, P = .0077). It was of greatest benefit in reducing hemorrhagic deaths if administered within the first hour of injury (OR = 0.68, 95% CI, 0.57-0.82, P < .001). Despite its antifibrinolytic effect, there was no increased risk of thromboembolic complications. Certainly, any patient bleeding sufficiently to get RBC and plasma transfusions should receive tranexamic acid, commenced immediately on arrival to hospital (or during transport if possible). The results of the CRASH-2 study were confirmed in an analysis of 896 consecutive United Kingdom and US military trauma admissions, of whom 293 had received tranexamic acid.47 The tranexamic acid cohort had a lower mortality rate (17.4 vs 23.9%, P = .03) despite having a higher injury severity score (25.2 vs 22.5, P < .001). It is estimated that if all eligible trauma patients worldwide received tranexamic acid within an hour of injury, 128 000 deaths might be averted.48
The WOMAN trial is comparing tranexamic acid with placebo in 15 000 women with postpartum hemorrhage to determine the effect on mortality, hysterectomy, and other morbidity.49 At the time of this writing (September 24, 2012), 4490 patients have been randomized. There is strong evidence that tranexamic acid reduces blood loss at the time of cesarean section.50 A small randomized trial in patients with postpartum hemorrhage showed that tranexamic acid reduced the risk of progression to severe postpartum hemorrhage and the risk of a 40 g/L decrease in hemoglobin.51 In addition, a recent meta-analysis update by the Cochrane group, including 252 trials and more than 25 000 patients, showed that antifibrinolytics reduced the risk of transfusion as a prophylactic measure before nonurgent cardiac and noncardiac surgery.52
Resuscitation with coagulation concentrates and goal-directed individualized care based on laboratory coagulation assessment will be discussed together, because the majority of publications describe the use of TEG to guide the use of fibrinogen and other coagulation concentrates. A Cochrane systematic review of 9 studies involving 776 patients concluded that there was an absence of evidence that TEG reduced morbidity or mortality in massively bleeding patients.53 In aortic, cardiovascular, and liver transplantation surgery, TEG reduces transfusion rates and risk of massive transfusion.54–57 A systematic review of fibrinogen concentrates58 detailed numerous fibrinogen studies. Three randomized trials were identified, 2 in the setting of cardiac surgery (1 adult and 1 pediatric trial) and the other in urologic surgery. The largest of these 3 trials included only 31 patients. Clearly, we do not have sufficiently designed randomized clinical trials to determine whether fibrinogen concentrates are an effective hemostatic agent in the setting of massive hemorrhage.
The highest-quality report (a nonrandomized comparative study) on the use of fibrinogen concentrates compared 80 trauma patients managed at the Salzburg trauma center (managed with fibrinogen concentrates and PCC based on rotational TEG) to 601 patients included in the German trauma registry (managed with a standard plasma approach).24 The injury severity scores were similar (35 for both groups). The fibrinogen-PCC group avoided RBC transfusion 29% of the time compared with only 3% in the German plasma cohort. Whether this is due to its impressive hemostatic effects or less hemodilution remains to be seen. The mortality rate was comparable (7.5% in the fibrinogen-PCC group and 10.0% in the plasma group, P = .69). We need a large, randomized trial comparing the standard plasma approach with TEG-guided concentrate therapy before it can be recommended as a standard approach. There are numerous benefits to the “concentrate” approach: less time needed to prepare these products (ie, they are not frozen and no ABO group is required), smaller volume (less hemodilution), and potentially fewer transfusion reactions (most notably, transfusion-related acute lung injury).
In addition, we do not know the level at which hypofibrinogenemia contributes to hemorrhage. A study from 1987 of 36 massively transfused patients (managed with modified whole blood: RBCs in cryosupernatant) concluded that coagulopathic type of bleeding was not observed until the fibrinogen level decreased below 1.0 g/L.59 The study included few patients, with only 14 of 36 having a fibrinogen level of less than 1.0 g/L. This is the study that has been repeatedly cited as the scientific basis for this cutoff. However, this threshold has recently been called into question, with some recommending a cutoff of 1.5-2.0 g/L.60,61 A study in patients undergoing cardiac surgery found that a low-normal preoperative baseline fibrinogen, even when within the normal range, increased the risk of postoperative transfusion.62 In addition, in the setting of postpartum hemorrhage, a fibrinogen level below 2.0 g/L at the commencement of bleeding increased the risk of severe postpartum hemorrhage.63 This threshold of 1.0 g/L in the massively bleeding patient must be questioned. We currently have no idea where the fibrinogen needs to be when managing a patient with marked blood loss. Further clinical studies are needed to determine the optimal fibrinogen level that should be incorporated into massive hemorrhage protocols.
The importance of having a protocol to optimize communication and coordination of care
The importance of the coordination of care for the massively bleeding patient cannot be overstated. Nunez et al published a comprehensive review of the steps required to build, implement, and monitor a massive transfusion protocol.64 The US military has been able to show that the implementation of clinical practice guidelines for massive transfusion can change care.37 In their review of 777 military patients, there were able to show an improvement in the temperature of patients on arrival, a reduction in crystalloid exposure, and an excellent compliance with their 1:1 protocol. In a review of 125 massive transfusion protocol activations in the setting of civilian trauma, Cotton et al found poor compliance (27% for all measures of performance)65 ; full compliance was associated with an improvement in survival (87% vs 45%, P < .001). Compliance may be a measure of better care by the clinical team, or it may simply reflect that it is easier to be compliant with the protocol with a less unstable patient, who by definition has a better chance of survival. A systematic review and meta-analysis of 7 observational, nonrandomized studies evaluating massive hemorrhage protocols was published recently.66 Of 1801 patients included in these studies, the use of a massive transfusion protocol was associated with a significant reduction in mortality, although no reduction in the number of RBC transfusions.
Follow-up of patients after severe hemorrhage
After hemostasis has been achieved, it is very important to consider whether the amount of hemorrhage seen was in keeping with the degree of injury or type of operative procedure. If the amount of hemorrhage was considerably greater than expected, consideration should be given to bringing the patient back several weeks after hemorrhage to determine whether there is an underlying bleeding disorder. In a case series of 317 women with severe postpartum hemorrhage, testing at 6-9 months postpartum detected 69 (22%) women with underlying hemostatic defects (28 with hypofibrinogenemia, 23 with low levels of VWF, and 18 with low factor XI).67 Knowing this history is extremely important, because a history of a postpartum hemorrhage is a risk factor for postpartum hemorrhage in a subsequent pregnancy.68
In the last decade, we have come to understand that the coagulation disturbance in the massively bleeding patient is more than just consumption and hemodilution. Activated protein C, hyperfibrinolysis, hypofibrinogenemia, and platelet dysfunction all play important roles. No doubt the introduction of novel anticoagulants will create additional challenges in the management of the massively bleeding patients. We have also come to the realization that the routine tests of coagulation are suboptimal in the management of massive hemorrhage. This has resulted in a resurgence of interest in TEG, although definite evidence on its utility in the massively bleeding patient is also lacking. Some European centers currently use rotational TEG to guide a predominately factor concentrate approach to massively bleeding patients (fibrinogen concentrates and PCCs), although no prospective trials have validated this approach. Although now commonly built into massive transfusion protocols worldwide, 1:1 plasma (or platelet) to RBC resuscitation is unproven. The prospective randomized trial PROPPR will hopefully clarify the role of plasma and platelets in the massively bleeding trauma patient. In terms of alternatives, there is no benefit of recombinant factor VIIa. In contrast, tranexamic acid should be included in all massive hemorrhage protocols due to the high level of evidence and minimal risk of adverse effects. Despite recent advancements in this field, the basic principles of patient care, such as promptly identifying the source of the bleeding and stopping it, should be strictly followed and not forgotten, particularly now that we know that the longer the patient remains in shock, the more coagulopathic s/he will become.
A consensus conference on massive transfusion,69 organized by the National Advisory Committee on Blood and Blood Products (Canada), provides a detailed list of recommendations for assisting with the development of an institutional massive hemorrhage policy. This document also highlights some of the unanswered questions to give guidance on where we need to go from here. What is obvious from the literature is that all centers need a predefined massive hemorrhage protocol to ensure rapid availability of blood and blood components, standardization of care, and clear communication. As new evidence becomes available, it will allow us to fine-tune these protocols to ensure the best possible outcomes for our patients.
Conflict-of-interest disclosure: S.R. received the Canadian Institutes of Health Research New Investigator Award in partnership with industry (Novo Nordisk). J.C. declares no competing financial interests. Off-label drug use: None disclosed.
Jeannie L. Callum, MD, FRCPC, Department of Clinical Pathology, Sunnybrook Health Sciences Centre, 2075 Bayview Ave, Rm B204, Toronto, ON, Canada M4N 3M5; Phone: 416-480-4045; Fax: 416-480-6035; e-mail: firstname.lastname@example.org.