Haemorrhage requiring massive transfusion remains a major cause of potentially preventable deaths. Trauma and massive transfusion are associated with coagulopathy secondary to tissue injury, hypoperfusion, dilution and consumption of clotting factors and platelets. Coagulopathy, together with hypothermia and acidosis, forms a ‘lethal’ triad (Lier et al, 2008). Also, in the last 10–15 years, there has been some paradigm shift regarding optimal resuscitation of bleeding trauma patients before definitive haemorrhage control is achieved. Aggressive fluid resuscitation increases blood pressure, reverses vasoconstriction, dislodges early formed thrombus, causes dilutional coagulopathy and metabolic acidosis and increases blood loss in experimental studies (Stern et al, 1993). According to previous guidelines (American Society of Anesthesiologists Task Force on Perioperative Blood Transfusion and Adjuvant Therapies, 2006), fresh frozen plasma (FFP) and platelets (PLT) should be administered only when a whole blood volume or more has been substituted and then according to conventional coagulation analyses. Yet, this strategy leads to dilutional coagulopathy and compromises haemostatic competence for the most severely bleeding patients (Lier et al, 2008). Instead, limiting fluid resuscitation with the goal of achieving a palpable radial pulse in patients has been advocated, whereas in patients with head injury a systolic blood pressure above 110 mmHg is recommended (Dries, 1996; Krausz, 2003; Soreide and Deakin, 2005).
The current transfusion guidelines have been challenged and the concept of haemostatic control resuscitation, i.e. supplementing large transfusions of red blood cells (RBC) with FFP and PLT to critically injured patients in an immediate and sustained manner is proposed (Johansson et al, 2005; Holcomb et al, 2007). The rationale for balanced administration of blood products is that it mimics the composition of circulating blood and, hence, transfusion of RBC, FFP and PLT in a unit-for-unit ratio is likely to both prevent and treat coagulopathy due to massive haemorrhage. Coagulopathy secondary to trauma is present already at the scene of the accident (Carroll et al. 2009) and consequently also pre-hospital resuscitation strategies, including administration of antifibrinolytics to the most severely bleeding and/or shocked trauma, influence both bleeding dynamics and further development of coagulopathy This article describes the clinical problems associated with haemorrhage and massive transfusion
Coagulopathy in massive haemorrhage
Dilution
The dilution of coagulation factors and platelets is an important cause of coagulopathy in massively transfused patients (Chappell et al, 2008). The Advanced Trauma Life Support guideline recommends aggressive crystalloid resuscitation: Initially by 2 L of lactated Ringer's solution (LR) and continued bolus infusions until completion of surgical repair (Kortbeek et al, 2008). The dilutional effects of such administration on coagulation competence are well described (Nielsen, 2005; Mittermayr et al, 2007) and this strategy provokes acidosis, formation of interstitial oedema with tissue swelling, impairment of the microcirculation, and hence compromised oxygenation (Knotzer et al, 2006).
Furthermore, synthetic colloid resuscitation fluids influence coagulation competence more profoundly than crystalloids. Hydroxyethyl starch (HES) cause efflux of plasma proteins from blood to the interstitial space, reduction in plasma concentration of coagulation factor VIII and von Willebrand factor (vWF), inhibition of platelet function and reduced interaction of activated FXIII with fibrin polymers (Mortier et al, 1997; Nielsen, 2005; Mittermayr et al, 2007). The latter effect causes slowly growing and weak clot formation less resistant to fibrinolysis. Colloid resuscitation fluids also induce hypocalcaemia, which further aggravates the coagulopathy.
The optimal pre-hospital fluid management was recently investigated in a randomized controlled clinical trial where trauma patients with hypovolemic shock were randomized to 250 mL of either 7.5% saline per 6% dextran 70 (hypertonic saline/dextran, (HSD)), 7.5% saline (hypertonic saline, (HS)), or 0.9% saline (normal saline, (NS)) administered by out-of-hospital providers. Initial resuscitation fluid treatment with either HS or HSD compared with NS, did not result in superior 28-day survival (HSD: 74.5% (0.1; 95%; HS: 73.0%; and NS: 74.4%, P=0.91 (Bulger et al, 2010).
Hypothermia
Hypothermia, defined as a core body temperature of 34–36 °C (mild); 32–34 °C (moderate) and below 32 °C (severe) is associated with risk of uncontrolled bleeding and death in massively transfused patients. Hypothermia induced coagulopathy is attributed to platelet dysfunction, reduced coagulation factor activity (significant below 33 °C) (Wolberg et al, 2004), and induction of fibrinolysis (Yoshihara et al, 1985). Platelet dysfunction and impaired coagulation enzyme activity are reversible with normalization of body temperature, highlighting the need to prevent and treat hypothermia aggressively (Romlin et al, 2007).
Acidosis
In massively transfused patients, acidosis is often induced by hypoperfusion and excess administration of ionic chloride, i.e. NaCl during resuscitation (Hess et al, 2008). Acidosis impairs almost all essential parts of the coagulation process: at pH < 7.4, platelets change their structure and shape, becoming spheres deprived of pseudopodia (Djaldetti et al, 1979). The activity of coagulation factor complexes on the cell surface is reduced and the resulting impaired thrombin generation is a major cause of coagulopathic bleeding. Furthermore, acidosis leads to increased degradation of fibrinogen (Martini and Holcomb, 2007) which further aggravates the coagulopathy. Administration of buffer solutions reverse acidosis but does not correct the coagulopathy, implying that pH affects more than protease activity.
Trauma
Trauma induces immediate activation of the coagulation system through upregulation of tissue factor (TF) expression and extensive thrombin generation. Tissue injury in association with extensive endothelial injury, massive soft tissue damage, and fat embolization from long bone fractures may be associated with consumption of coagulation factors and platelets and, hence coagulopathy, which together with dilution from intravenous fluid therapy, hypothermia, and metabolic acidosis contribute to the ‘bloody viscous cycle’ (Hess and Lawson, 2006; Levi, 2007).
Brohi et al (Brohi et al, 2003; 2007a; 2007b; 2008) described an early ‘endogenous’ coagulopathy in trauma patients not attributed to the abovementioned causes. According to their paradigm, shock and hypoperfusion are the key drivers of acute post-traumatic coagulopathy through widespread activation of the anticoagulant and fibrinolytic Thrombomodulin (TM)-protein C (PC) pathways.
Once activated through a thrombin-dependent reaction involving TM and possibly the endothelial protein C receptor (EPCR), activated PC (aPC) exerts its anticoagulant effects by irreversibly inactivating FVa and FVIIIa, by limiting continued thrombin production, and by enhancing fibrinolysis of already formed clots through inhibition of PAI-1, a key inhibitor of tPA. Consequently, acute post-traumatic coagulopathy is characterized by systemic anticoagulation in conjunction with hyperfibrinolysis (Brohi et al, 2007a).
Monitoring haemostasis
Introduction of the cell-based model of haemostasis emphasizes the role of platelets for intact thrombin generation and highlights the importance of the dynamics of thrombin generation, influencing the quality and stability of the thrombus formed (Roberts et al, 2006). Consequently, haemostatic assays performed on plasma, such as activated partial thromboplastin time (APTT) and prothrombin time (PT), are of limited value (Fries et al, 2009) and do not correlate with clinically relevant coagulopathies (Murray et al, 1999; Segal and Dzik, 2005).
Instead, employing a whole blood assay, such as TEG that records the viscoelastic changes during clot formation and subsequent lysis, is preferable. The TEG reports (Figure 1): reaction time (R), the latency from the time at which the blood is placed in the cup until the clot begins to form; angle (α), the progressive increase in clot strength; the maximum amplitude (MA), the maximal clot strength; and clot lysis (Ly) the reduction in amplitude at a specified time point after reaching MA (Salooja and Perry, 2001; Luddington, 2005; Ganter and Hofer, 2008; Johansson et al, 2009).
Thrombus generation, as evaluated by TEG, correlates with thrombin generation kinetics (Johansson et al, 2008) explaining why TEG identify clinically relevant coagulopathies (Reikvam et al, 2009). Accordingly, reduced clot stability correlates with clinical bleeding conditions as demonstrated by Plotkin et al (2008) who, in patients with a penetrating trauma, reported TEG to be an accurate indicator of the blood product requirements. Furthermore, TEG is the gold standard for identifying hypercoagulability (Park et al, 2009) and hyperfibrinolysis (Levrat et al, 2008), the latter a significant cause of bleeding in major trauma, ischemia/reperfusion injury and obstetric calamities (Rugeri et al, 2007; Levrat et al, 2008; Schochl et al, 2009).
There are developed treatment algorithms for how coagulopathy is identified and treated based on TEG that are validated in patients with ongoing bleeding (Johansson and Stensballe, 2009). More than 25 studies encompassing more than 4500 patients have evaluated TEG vs conventional coagulation assays on bleeding and transfusion requirements in patients undergoing cardiac, liver, vascular, or trauma surgery and in patients requiring massive transfusion. These studies demonstrate the superiority of TEG in predicting the need for blood transfusion, and the algorithm reduces the transfusion requirements and the need to re-do surgery in contrast to treatment based on a plasma-based coagulation assay (Johansson et al, 2009). The updated 2010 European guideline for management of bleeding following major trauma now recommends that ‘Thrombelastography may assist in characterizing coagulopathy and guiding haemostatic therapy’ (Rossaint et al, 2010).
Administration of blood products
Red blood cells
In response to haemorrhage, lowered hematocrit contributes to coagulopathy since erythrocytes promote marginalization of platelets so the platelet concentrations along the endothelium remains almost seven times that of the average blood concentration (Uijttewaal et al, 1993). In addition, erythrocytes support thrombin generation through exposure of procoagulant membrane phospholipids (Peyrou et al, 1999; Lier et al, 2008), they activate platelets by liberating ADP (Valles et al, 1991; Joist et al, 1998) emphasizing that in vivo thrombus formation is a multicellular event (Roberts et al, 2006; Monroe and Hoffman, 2006). Yet, the optimal haematocrit for platelet–vessel wall interactions remains unknown but it may be as high as 35% and, consequently, well above the level needed for oxygen delivery (Hardy et al, 2004).
Fresh frozen plasma
It remains controversial when, and in what dose, plasma should be transfused to massively bleeding patients, although retrospective cohort studies report a survival benefit in patients receiving a high FFP to RBC ratio as measured in units (Stinger et al, 2008; Sperry et al, 2008; Maegele et al, 2008; Duchesne et al, 2008; Kashuk et al, 2008; Holcomb et al, 2008; Snyder et al, 2009; Teixeira et al, 2009; Zink et al, 2009; Wafaisade et al, 2011). Duchesne et al (2008) found in a multivariate analysis that a FFP:RBC ratio of 1:4 was consistent with increased mortality compared to a FFP:RBC ratio of 1:1. Thus, the level of coagulation factors and fibrinogen is important for haemostasis and, ultimately survival.
The optimal ratio of FFP to RBCs remains to be established although collectively the data indicate that a FFP:RBC ratio greater than 1:2 is associated with improved survival compared to one lower than 1:2 (Duchesne et al, 2008; Kashuk et al, 2008; Holcomb et al, 2008; Snyder et al, 2009; Teixeira et al, 2009; Wafaisade et al, 2011). This is further supported by a recent meta-analysis reporting that in patients undergoing massive transfusion, high FFP to RBC ratios was associated with a significant reduction in the risk of death (odds ratio (OR) 0.38 (95%CI 0.24-0.60) and multiorgan failure (OR 0.40 (95%CI 0.26-0.60) (Murad et al, 2010).
Platelets
Platelets are also pivotal for haemostasis (Roberts et al, 2006; Monroe and Hoffman, 2006) and several retrospective studies report an association between thrombocytopenia and postoperative bleeding and mortality (Adam et al, 2003; Johansson et al, 2005; Johansson et al, 2007). Furthermore, thrombocytopenia upon arrival at the ICU after surgery is consistently associated with poor outcome (Stephan et al, 1999). Several observational studies report on the effects of PLT transfusion (Stinger et al, 2008; Holcomb et al, 2008; Zink et al, 2009; Perkins et al, 2009; Inaba et al, 2010b) and demonstrate improved survival in patients receiving (most) platelets. Holcomb et al (2008) found that the highest survival was established in patients who received both a high PLT:RBC and a high FFP:RBC ratio. Inaba et al recently reported from a retrospective study of massively transfused patients that as the apheresis platelet to RBC ratio increased, a stepwise improvement in survival was seen and a high apheresis PLT:RBC ratio was independently associated with improved survival (Inaba et al, 2010a). Intriguingly, Salim et al (2009) found that posttraumatic thrombocytosis, which manifested in 19% of the patients, were associated with improved survival (18% vs 4%).
Massive transfusion protocols
Cotton et al (2008) implemented a trauma exsanguination protocol (TEP) involving 10 RBC, 4 FFP and 2 apheresis PLT for trauma patients and used it to evaluate 211 trauma patients of who 94 received TEP and 117 were historic controls. The TEP patients received more RBC (16 vs 11), FFP (8 vs 4), and PLT (2 vs 1) intraoperatively than the controls and displayed lower 30-day mortality (51% vs 66%). After controlling for age, sex, mechanism of injury, Trauma and Injury Severity Score (TRISS), and 24-hour blood product usage, a 74% reduction in the odds ratio of mortality was found among patients in the TEP group.
We evaluated 832 consecutive mixed massively transfused patients two years prior to (2002–2003) and two years after (2005–2006) implementation of haemostatic control resuscitation (Johansson and Stensballe, 2009). The concept involved pre-emptive use of PLT and FFP organized into transfusion packages (5 RBC, 5 pre-thawed FFP and 2 PLT; each a pool of 4 buffy coat platelets) to be administered to patients with uncontrollable bleeding. When haemodynamic control was established, the transfusion therapy was directed by whole blood TEG analyses according to a protocol (Table 1). The FFP:RBC ratio in the intervention group, as expressed in units, was 1:1.3 vs 1:1.6 in the controls. Compared to the controls, the intervention group received more PLT within 24-hours of admission (mean (SD) 5.0 (4.2) vs 1.7 (2.0)), exhibited a smaller decrease in platelet count during the bleeding episode (92 (81) vs 120 (101)×109/l) and a reduction in 30-day mortality (20% vs 31% in controls). In alignment with this, Duchesne et al reported that in trauma patients undergoing damage control laparotomy introduction of damage control resuscitation encompassing early administration of plasma and platelets together with RBC was associated with improved 30-day survival, (74% vs 55%, P<0.009) when compared to patients treated with conventional resuscitation efforts (Duchesne et al, 2010). Furthermore, Cotton et al (2009) reported on the cohort of 264 trauma patients described above, that not only was 30-day survival higher in the TEP group compared to the controls, but the incidence of pneumonia, pulmonary failure and abdominal compartment syndrome was lower in the TEP patients. Also, the incidence of sepsis or septic shock and multi-organ failure was lower in TEP patients. Although the TEP group received more blood products intraoperatively, the 24-hour transfusion requirements were lower than in controls, supporting that early and aggressive administration of plasma and platelets improves haemostasis.
| TEG Parameters | Coagulopathy | Intervention |
|---|---|---|
| R 10–14 | Coagulation factors ↓ | FFP 10–20 ml/kg, alternatively Cryoprecipitate pool 3 ml/kg |
| R>14 min | Coagulation factors ↓↓ | FFP 10–20 ml/kg, alternatively Cryoprecipitate pool 5 ml/kg |
| MA 45–49mm | Platelets and/or fibrinogen ↓ | PLT 5 ml/kg |
| MA <45 mm | Platelets and/or fibrinogen ↓↓ | PLT 10 ml/kg |
| MA <49 mm and MAFF <14 mm | Fibrinogen ↓ or ↓↓ | FFP 20–30 ml/kg or Fibrinogen conc. 25-50 mg/kg or Cryoprecipitate pool 5 ml/kg |
| MA <49 mm and MAFF >14 mm | Platelets ↓ or ↓↓ | PLT 5–10 ml/kg |
| Ly30 >8% | Primary hyperfibrinolysis | TXA 1–2 g (children 10–20 mg/kg) |
| Ly30 >8% and Angle and/or MA ↑↑ | Reactive hyperfibrinolysis | TXA contraindicated |
R=reactiontime, alphaangle=clotdynamics, MA=Maximal Amplitude, MAFF=Functional Fibrinogen MA (rapid evaluation of plasma fibrinogen function), Ly30=lysis in percent 30 min after MA is reached, FFP=freshfrozen
Haemostatic agents
Antifibrinolytics
Hyperfibrinolysis contribute significantly to coagulopathy and antifibrinolytics agents reduce the blood loss in patients with both normal and exaggerated fibrinolytic responses to surgery by preventing plasmin(ogen) from binding to fibrin and by preventing plasmin degradation of platelet glycoprotein Ib receptors (Porte and Leebeek, 2002; Henry et al, 2007). In a recently published placebo controlled randomized study (CRASH-2) including 20 211 adult trauma patients, tranexamic acid as compared to placebo significantly decreased all-cause mortality from 16.0%–14.5%, P=0.0035 (Shakur et al, 2010). We recommend monitoring of haemostasis with TEG to identify hyperfibrinolytic states in trauma patients (Kaufmann et al, 1997; Rugeri et al, 2007; Levrat et al, 2008; Carroll et al, 2009; Schochl et al, 2009) and, consequently, targeted treatment with antifibrinolytic agents.
Recombinant factor VIIa
Recombinant factor VIIa (rFVIIa) acts in pharmacological doses by enhancing thrombin generation on the activated platelets independent of factor VIII and IX and is currently approved for episodes of severe haemorrhage or perioperative management of bleeding in patients with congenital factor VII deficiency and haemophilia A or B with inhibitors (Hedner, 2008; Johansson and Ostrowski, 2010). Since the first report of rFVIIa administered to a trauma victim in 1999 (Kenet et al, 1999), there has been off-label use of rFVIIa for the management of various non-haemophilic bleeding conditions although none of the published RCT report of improved effect on mortality in rFVIIa treated patients when compared to placebo treated patients (Johansson, 2008).
In a recent review reporting on the rate of thromboembolic events in all published randomized placebo controlled trials of rFVIIa use (35 trials, 4 468 subjects) (Levi et al, 2010), the rates of arterial thromboembolic events among all subjects were higher among those who received rFVIIa than among those who received placebo (5.5% vs 3.2%, P=0.003), with particularly increased arterial thromboembolic rates in subjects older than 65 years. In 2008, Novo Nordisk discontinued a phase 3 clinical trial with NovoSeven® for the treatment of bleeding in patients with severe trauma and declared that they would not pursue trauma as an indication for the use of rFVIIa (Drugs Information Online, 2008).
Yet, in patients with a massive uncontrolled life threatening blood loss not responding to conventional treatment, administration of rFVIIa may be considered for the damage control strategy, although evidence for such an approach is lacking (Spinella et al, 2008; Bartal and Yitzhak, 2009). Furthermore, recent data from the CONTROL trial, a phase 3 randomized clinical trial evaluating efficacy and safety of rFVIIa as an adjunct to direct hemostasis in major trauma (Hauser et al, 2010), rFVIIa did not change mortality in patients with blunt (11.0% (rFVIIa) vs 10.7% (placebo)) or penetrating (18.2% (rFVIIa) vs 13.2% (placebo)) trauma despite a significant reduction in RBC and total allogenic units in patients with blunt injuries and significantly fewer patients requiring massive RBC transfusion in patients with penetrating injuries.
Importantly, administration of rFVIIa should be preceded by administration of platelets and fibrinogen, and if possible treatment of hypothermia and acidosis, to ensure a possibility for rFVIIa to act.
Fibrinogen
Conversion of sufficient amounts of fibrinogen to fibrin is a prerequisite for clot formation and reduction in the circulating level of fibrinogen due to consumption (Hess and Lawson, 2006) and/or dilution by resuscitation fluids (Hiippala et al, 1995) induces coagulopathy (Nielsen et al, 2005). Fibrinogen is the first haemostatic component that declines to pathologically low levels following trauma and/or haemodilution (Hiippala et al, 1995). It is therefore important to maintain an adequate fibrinogen level when continued bleeding is bridged by saline and/or colloid infusion.
Recent data indicate that coagulopathy induced by synthetic colloids may be reversed by the administration of fibrinogen concentrate (Fenger-Eriksen et al, 2009). Similarly, a high fibrinogen to RBC ratio improved survival in combat casualties requiring massive transfusion (Stinger et al, 2008). In thrombocytopenic pigs, fibrinogen concentrate administration improved clot formation and survival better than PLT (Velik-Salchner et al, 2007), emphasizing that fibrinogen exerts clinically relevant haemostatic effect also during thrombocytopenia.
Prothrombin complex concentrate
The PCC encompasses coagulation factors II, VII, IX and X that all are essential for thrombin generation. Administration of PCC is indicated for the treatment of congenital coagulation disorders and to reverse oral administered anticoagulation by vitamin K antagonists (Pabinger-Fasching, 2008), whereas experience with treatment of massive bleeding with PCC is lacking.
Pre-hospital interventions
In a recent systematic review evaluating haemostatic dressings, HemCon and QuikClot were reported to augment the haemostatic capabilities of the military first aid responder, though newer products demonstrate potential to be more effective and should be considered as replacements for current in service systems. Of particular importance, it was concluded that these products could also be used for civilian pre-hospital care.(Granville-Chapman et al, 2010)
In rare situations tourniquet application will be necessary and lifesaving in the civilian pre-hospital setting. Tourniquets are no longer only considered a ‘last resort’ device. Practitioners should familiarize themselves with this simple piece of equipment and be prepared to use it in appropriate cases and thereby arrest life-threatening external haemorrhage from limb injury.(Kragh Jr et al, 2009)
Conclusion
Viscoelastic whole blood assays, such as TEG, are advantageous for identifying coagulopathy, and guide ongoing transfusion therapy. From the result of these assays, implementation of a haemostatic control resuscitation strategy to massively bleeding patients seems both reasonable and lifesaving although data from prospective randomized controlled trials are lacking. Until definite proof from such trials is available, retrospective data support a shift in transfusion medicine in regard to early and aggressive administration of plasma and platelets.