For 30 years, the Advanced Trauma Life Support course of the American College of Surgeons taught that coagulopathy was a late consequence of resuscitation of injury. The recognition of trauma-induced coagulopathy overturns that medical myth and creates a rationale for procoagulant resuscitation. Analysis of the composition of currently available blood components allows prediction of the upper limits of achievable coagulation activity, keeping in mind that oxygen transport must be maintained simultaneously. RBCs, plasma, and platelets given in a 1:1:1 unit ratio results in a hematocrit of 29%, plasma concentration of 62%, and platelet count of 90 000 in the administered resuscitation fluid. Additional amounts of any 1 component dilute the other 2 and any other fluids given dilute all 3. In vivo recovery of stored RBCs is ∼ 90% and that of platelets ∼ 60% at the mean age at which such products are given to trauma patients. This means that useful concentrations of the administered products are a hematocrit of 26%, a plasma coagulation factor activity of 62% equivalent to an international normalized ratio of ∼ 1.2, and a platelet count of 54 000. This means there is essentially no good way to give blood products for resuscitation of trauma-induced coagulopathy other than 1:1:1. Because 50% of trauma patients admitted alive to an academic-level 1 trauma center who will die of uncontrolled hemorrhage will be dead in 2 hours, the trauma system must be prepared to deliver plasma- and platelet-based resuscitation at all times.
Who needs resuscitation of trauma-induced coagulopathy?
The fraction of injured patients exhibiting trauma-induced coagulopathy (TIC) increases as injury severity increases to involve at least a quarter of all severely or profoundly injured patients.1 Endothelial disruption is the driving force for the initial activation of coagulation, and shock and the activation of protein C are the most important drivers of coagulation factor consumption and the loss of inhibition of fibrinolysis.2,3 Hypothermia profoundly reduces platelet activation and, along with acidosis, reduces coagulation factor enzyme activity.2 Genetic clotting defects; acquired liver, blood, and vascular disease; and medications can all contribute in multiplicative ways to making blood clotting worse in the injured.4
Because coagulopathy is uncommon in uninjured individuals, its prevalence in young people is essentially the 1% prevalence of von Willebrand disease.5 This prevalence increases with age as acquired disease, most frequently cirrhosis, and the use of medications such as warfarin and dual antiplatelet therapy become more common. As injury severity increases, coagulopathy becomes more common, rising from several percent in the moderately injured with injury severity scores ranging from 9 to 16, to > 10% in severely injured individuals, with injury severity scores of 17 to 25 and ∼ 40% in those with profound injury as measured by an injury severity scores > 25.6
The combination of severe injury and coagulopathy is associated with excess mortality. This excess mortality is proportional to both the injury severity and the extent of coagulopathy, but if one defines the coagulopathy as a prolonged prothrombin time (PT) or partial thromboplastin time (PTT), the excess mortality is of the order of 4-fold.1 Low numbers of platelets and reduced plasma fibrinogen concentrations are also associated with equivalent excess mortality.4
Mortality from severe injury and coagulopathy occurs quickly. Half of all injured civilians who die from their injuries die before reaching the hospital.7 Such patients typically have profound neurologic injury or major vascular disruption. Among soldiers who die of battlefield injuries, the fraction dying before they reach surgical care is even greater, approaching 85%.8,9 In both groups, the largest proportion of those who die but are judged to have had a potentially salvageable injury had uncontrolled hemorrhage as their mechanism of death.10 When such patients arrive at the hospital alive, their situation is urgent. Among 68 000 patients who arrived directly from the scene of injury to the University of Maryland's Trauma Center in Baltimore between 1996 and 2008, the mean time to death for those who died of uncontrolled hemorrhage was 2 hours.11 The curve describing the rapidity with which these patients with uncontrolled major hemorrhage die—30% in the first hour, 50% by 2 hours, 63% by 3 hours, 80% by 6 hours, and 90% by 12 hours—means that lifesaving interventions must be given well before these times to have a chance of being effective. This is the metaphoric “golden hour.”12
Resuscitation is one of the goals trauma teams try to accomplish within that golden hour. The objectives of resuscitation are to restore cardiac output, oxygen delivery, and blood coagulation. It was not always so. For 3 decades, the Advanced Trauma Live Support (ATLS) course of the Committee on Trauma of the American College of Surgeons has taught that cardiac output can be maintained with crystalloid fluids, that oxygen delivery can be supported with RBCs, and that abnormalities of blood coagulation occur late in resuscitation and can be treated based on abnormalities of readily available laboratory tests.13 We now know that crystalloid fluids maintain intravascular volume poorly and cause tissue edema, compartment syndromes, and inflammation.14 Leukocyte-reduced RBCs in additive solution contain no platelets and only small amounts of plasma. We know that acute coagulopathies of trauma and shock exist, but conventional laboratory tests of coagulation are rarely available quickly, do not address fibrinolysis, and are generally interpreted without reference to values associated with excess mortality in trauma patients.
How to manage resuscitation of TIC with blood components
Modern resuscitation of the most seriously injured involves giving less crystalloid and more plasma. However, there are many barriers to giving more plasma. Plasma is generally stored frozen and takes time to thaw. Universal donor AB plasma is a rare commodity, available from only 4% of donors and needed for AB patients and babies. Thawing plasma for patients who have a 30% chance of being dead in an hour raises the specter of wasting AB plasma. There are dicta from the highest levels of our profession that “plasma should not be used as a volume expander.”15 There are guidelines that plasma should not be issued to patients with an international normalized ratio < 1.5 times normal.16 There are prejudices from proponents of evidence-based medicine that the lack of level 1 evidence favoring early plasma transfusion is the lack of any real evidence at all. Overcoming these barriers to early plasma transfusion requires a commitment to making thawed plasma available early after the arrival of injured patients.
Making thawed plasma available quickly is possible in several ways. Rapid thawing devices such as microwave blood warmers are available that will thaw 2 units of plasma in ∼ 2 minutes. Using several such devices allows the thawing and issue of 4 to 8 units of plasma in 5 to 10 minutes. Prethawed plasma can be kept for 5 days, so a transfusion service that uses 20 units of plasma a day can maintain 4 units thawed without much wastage, especially if they are willing to rotate stock and issue the unused units as universal donor plasma. Prethawed plasma can be preissued to trauma center self-service blood refrigerators and matched to patients after use. Never-frozen “liquid” plasma can be obtained from blood donor centers and kept for periods of up to 26 days. The supply of AB plasma can be expanded by collecting 600 mL units by apheresis from specially recruited donors as frequently as 24 times a year, and the definition of universal donor plasma can be expanded to include units of A plasma with low hemagglutinin titers of anti-B (eg, < 1:250).17,18 Finally, lyophilized plasma is being developed that can be reconstituted quickly at the time and site of use and is already available in Australia, Belgium, France, and Germany.19 Qualitative differences between these products are reported but largely unknown.
The coagulation activities of plasma and concentrations of platelets required to prevent coagulopathy in various surgical situations are largely a matter of opinion, being generally formulated as expert group guidelines because no large clinical trials exist in this area. The American Society of Anesthesia and the College of American Pathologists state that major elective surgery can be performed along anatomic planes when the PT and PTT are not > 1.5 times normal and platelets are > 50 × 109/L.20,21 The higher demands for bleeding control in neurologic and ophthalmic surgery have led to recommendations that the PT and PTT should not exceed 1.3 times normal and that the platelet count should be > 100 × 109/L. Conversely, less invasive procedures such as chest tubes and central lines can be safely performed with the PT or PTT prolonged to 1.8 × normal and the platelet count as low as 30 × 109/L according to the Massachusetts General Hospital guidelines, and some minimally invasive procedures such as fine needle biopsy of breast and thyroid, placement of peripherally introduced central catheters, and BM biopsy appear to be safe at any platelet count.22 The overall pattern of these widely observed guidelines is that greater injury and critical organ involvement justifies less restrictive transfusion triggers. With the most severe injuries, the anecdotal experience and clinical series of our military surgeons using fresh whole blood suggests the value of attempting to restore coagulation completely.23,24
A similar experience exists with the most common single coagulation factor deficiency, hemophilia A. Severe, moderate, and mild disease are associated with factor VIII concentrations of < 1%, 2% to 5%, and 6% to 30%, respectively, where severe disease is associated with spontaneous bleeding and mild disease is typically problematic only after surgery or trauma.25 Treatment with 50% restoration of factor VIII activity for 3 days is usually be adequate for soft tissue and joint bleeds, but critical injury requires full restoration of 100% activity for 10 days to prevent rebleeding.
There is no way to achieve 100% of normal coagulation activity with transfusion of current blood products.26 This fact is a direct result of the way modern blood components are manufactured. We tend to think of blood components as “concentrates,” in that the RBCs in “packed RBCs” have a hematocrit of 55% and the platelets in platelet concentrates have a concentration 4 times higher than in normal whole blood. However, clinical plasma is diluted to 80% of normal concentrations by the addition of a 14% volume-to-volume ratio of citrate anticoagulant to whole blood during blood collection. When these components are mixed together in patients during care, the resulting mixtures are generally dilute in all of the components.
The extent to which the composition of individual blood components must determine the ratios of components given during massive transfusion can be understood by consideration of the exact compositions of clinical RBC, platelet, and plasma products. In modern whole blood processing, 450 mL, the old pint, or 500 mL, the new half-liter, of whole blood are collected into 63 or 70 mL, respectively, of citrate, phosphate, and dextrose anticoagulant. This dilutes the suspending plasma but leaves the RBC volume unchanged. The whole blood in anticoagulant is then gently spun, precipitating the RBCs, and most of the platelet-rich plasma supernatant is expressed off the top of the collection bag into an attached set of component separation bags.
Assuming that the donor has a hematocrit of 40%, there are 200 mL of RBCs in the 500 mL donation and the RBC bag contains all of these along with ∼ 50 mL of the plasma and anticoagulant mixture at a hematocrit of 80%. RBCs in this format store poorly and flow like molasses, so 110 mL of RBC nutrient additive solution is added. The resulting suspension now has the 200 mL of RBCs in 40 mL of plasma, 10 mL of anticoagulant, and 110 mL of nutrient additive solution for a 360 mL volume with hematocrit of 56%. Loss of 35 mL of this mixture in a leukocyte reduction filter results in the final typical 180 mL of RBCs in 325 mL of total volume.
The 320 mL of expressed platelet-rich plasma is then centrifuged again to sediment the platelets and the now platelet-poor plasma supernatant is expressed off the top of the platelet concentrate below. Typically, the division is 50 mL for the platelet concentrate and 270 mL for the platelet-poor plasma. Remembering that the platelets in half a liter of blood occupy only 1 mL, that many platelets are retained in the RBC mass and then lost in the leukocyte reduction filter, and that the plasma in the platelets and plasma bags is only 80% plasma, one can calculate the results of mixing various combinations of blood components.
The simplest example is the attempt to reconstitute whole blood by mixing one unit of RBCs with one unit of plasma and one unit of platelets, a 1:1:1 ratio. In this mixture, there are 180 mL of RBCs in a 645 mL total volume for a hematocrit of 28%. The 465 mL of remaining fluid contains only 288 mL of plasma for a coagulation factor concentration of 62% of that in normal plasma. Platelet losses in the leukocyte reduction filters reduce the final count of platelets in a whole blood derived unit of platelets to 600 to 800 × 109 or their final concentration in the mixture to 90 to 120 × 109/L. The in vivo recovery of radioactively labeled fresh platelets is ∼ 50% in storage studies, and 5-day stored platelets have a recovery 30% less than that.27 This means that the effective return of stored platelets is 20% of the starting number, their individual function is degraded, and the plasma they are suspended in will have lost some activity too. RBCs also lose viability during storage, so that only 90% of transfused RBCs actually circulate, reducing the effective concentration from 28% to 25%.28 The end result of these simple facts is that even if one is trying to resuscitate an injured patient with uncontrolled bleeding by giving nothing but blood products in a 1:1:1 ratio, the effective concentrations in the total fluid given is a hematocrit of 25%, a plasma concentration of 62%, and platelet concentration of ∼ 50 × 109/L. One is essentially at the transfusion trigger for all 3 components. Adding more of one component merely dilutes the other 2 and adding any other fluid dilutes all 3.
Some have argued that giving other ratios might be just as good and have suggested adding 2 units of RBC to every unit of plasma and platelets as an obvious choice (2:1:1).29 Under these circumstances, the concentration of RBCs in the administered fluid goes up and that of plasma and platelets goes down. The exact numbers are a hematocrit of 37%, a plasma concentration of 52%, and a platelet concentration of 62 to 93 × 109/L. Again, because in vivo recoveries of the platelets are only 50%, the effective administered platelet concentrations are 31 to 46 × 109/L, values well below the accepted transfusion trigger and with no potential to raise platelet concentrations above it.
It is possible to take advantage of the small differences in available components, such as choosing apheresis platelet concentrates that are collected in doubly concentrated anticoagulant so their plasma concentration is 93% rather than the 80% of whole blood derived units and using apheresis platelet units with higher platelet counts such as those with > 4 × 1011 platelets/unit rather than the legal minimum of 3.0. The advantages are marginal and can be lost if a blood bank issues one of the new units of platelets in additive solution containing 200 mL of saline and 100 mL of plasma rather than the conventional units with 300 mL of plasma.
It is important to recognize that even if volume replacement with a 1:1:1 ratio of RBCs, plasma, and platelets is the optimal way to resuscitate patients with uncontrolled hemorrhage and TIC, such patients represent only a small fraction, perhaps 2% of those admitted to a trauma center. The use of diagnostic algorithms to increase diagnostic specificity while maintaining sensitivity will reduce plasma and platelet waste and, in several large consecutive series, has been shown to reduce total blood use. The most impressive example comes from the National Institutes of Health's trauma glue grant consortium, where earlier use of plasma and platelets reduced mortality and total blood use.30 Adjuvants to the resuscitation of TIC include tranexamic acid and, rarely, rVIIa. Tranexamic acid is a lysine analog and inhibitor of plasmin-medicated fibrinolysis. It provided a significant survival benefit in the large CRASH II trial when given in the first 3 hours after injury.31 Administered rVIIa can replace and augment endogenous factor VII consumed with the massive exposure of tissue factor. It has failed to improve outcome in 2 large clinical trials when given late in dilutional coagulopathy.32,33
Resuscitation of TIC with clotting factor concentrates
An alternative form of treatment for TIC is the administration of pharmaceutically prepared plasma-derivative coagulation factor concentrates. The idea is that the intrinsic coagulation system can be largely reconstituted in the injured patient with 4-factor prothrombin complex concentrates (PCCs) and fibrinogen concentrate.34 The 4-factor PCCs provide factors VII, IX, X, and II and recently became available in the United States. Factor VIII, produced by the endothelium, is rarely deficient, and factor V is present in platelet alpha granules. Fibrinogen concentrates contain enough factor XIII to cross-link fibrinogen normally.
In Europe, these materials are available and, despite their greater cost, are preferred because of their greater presumed safety and are provided within some national medical systems. The freeze-dried proteins can be kept on the anesthetist's cart, reconstituted quickly, and given early. Austrian researchers claim that their use has improved outcomes and cut blood use in the Vienna trauma center by half.35 The concept is interesting, but the data are still scant. Randomized clinical trials are needed. The products have only recently become available in the United States and for other indications. The clinical consequences for the severely injured of low concentrations of VWF and other plasma proteins not present in the PCCs are not known. There is a general need for better clinical blood products and clinical effectiveness research.
Making resuscitation of TIC a trauma center quality issue
The American College of Surgeons Committee on Trauma runs the Trauma Quality Improvement Project (TQIP).36 One of its goals for this year is to try to improve the availability and timeliness of plasma-based resuscitation, not only for patients seen in level 1 trauma centers, but also for patients cared for in less intensively resourced settings.37 This means that, as a condition of accreditation, we will all need a plan to recognize and treat TIC.
At this time, the TQIP project centers on expanding massive transfusion protocols from just providing universal donor RBCs to creating systems that can quickly mobilize and issue balanced amounts of RBCs, plasma, and platelets. Five years ago, most hospitals did not have massive transfusion protocols. Now they are a requirement for accreditation. In the future, the standards will be more exacting. These new standards will drive changes in your transfusion service that will need to be carefully considered to optimize patient care and minimize blood component waste. It is hoped hematologists will support this effort in practice and through their transfusion services, transfusion committees, and local blood provider networks.
Conflict-of-interest disclosure: The author is on the board of directors or an advisory committee for CSL Behring. Off-label drug use: fibrinogen concentrate and 4-factor prothrombin complex concentrate.
John R. Hess, Box 359743, Blood Bank, 325 Ninth Ave North, Seattle, WA 98104; Phone: 206-897-6121; e-mail: firstname.lastname@example.org.