Successful outcome from damage control surgery (DCS) depends as much on elements of resuscitation and non-operative management as on details of the procedure itself. The early management of patients in haemorrhagic shock has undergone substantial revision in the past decade and is now known as ‘haemostatic resuscitation’.
An updated literature review describing the anaesthetic and resuscitative management of patients with active, ongoing traumatic haemorrhage was distilled to present the current knowledge of the pathophysiology, recommended treatments and areas of active controversy.
Current practice in military and civilian trauma centres is described, along with the degree of evidence in support of clinical decisions. Resuscitation of patients with ongoing traumatic haemorrhage has changed substantially in the past two decades. Optimal management now includes deliberate hypotension to minimize blood loss, early use of blood products (especially plasma) and administration of antifibrinolytic therapy. Areas of debate include the role of clotting factor concentrates and depth of anaesthesia.
Damage control surgery (DCS) advocates focused efforts to correct life-threatening anatomical insults after major trauma, while deferring repair of less severe injuries. DCS is based on the premise that time spent in the operating room (OR) is dangerous to haemodynamically unstable patients. It is difficult to challenge this assertion, because observational studies over the past three decades, before and after the widespread application of damage control techniques, have shown a survival benefit1. Some of the presumed benefit of DCS arises from mechanical elements of surgery, whereas reduced surgical time means less exposure to anaesthesia-guided fluid resuscitation, transfusion and medications, and this is presumed to be a further benefit of DCS.
The overarching goal of DCS is to stop haemorrhage, and the resuscitation strategy plays an important role in this process. In part, haemostasis is facilitated by avoidance of treatments that are known to exacerbate coagulopathy and rebleeding. A second, more challenging, goal is to create physiological conditions that actively facilitate clotting, minimize tissue ischaemia, and enhance humoral and endothelial integrity.
This article outlines the contributing pathophysiological factors, and suggests ways in which the non-anatomical components of resuscitation can be optimized. A case can be made that appropriate administration of fluids and medications to the unstable injured patient is at least as important to long-term outcomes as rapid anatomical control of bleeding vessels, although perhaps less widely achieved. Consequently, this review discusses ‘haemostatic resuscitation’ for injured patients in an unstable condition, and outlines accepted facts, current controversies and future possibilities for anaesthetic management during DCS.
Pathophysiology of shock and systemic response to injury
Shock, ‘a rude unhinging of the machinery of life’ (Samuel Gross, 1862), has been recognized and studied for more than a century2. Shock arising from trauma was once thought to be a simple model of pathophysiology: the loss of circulating blood volume led to reduced cardiac output and vasoconstriction of less essential organs. The reduced blood flow led to cellular ischaemia and eventual necrosis. Eventually cell death led to organ system failure and death of the patient. In many older publications, the terms ‘traumatic shock’ and ‘haemorrhagic shock’ are used interchangeably.
In recent years the pathophysiology of shock after trauma has been recognized as considerably more complex. Direct tissue injury provokes a local response from the affected cells, which in turn invokes a systemic reaction that may be greatly amplified3, 4. Several compounds are released from injured cells. Some of these, such as the troponins used in cardiac diagnostics, are present in the plasma as a simple consequence of membrane disruption. Others, such as tissue factor, are expressed deliberately as part of the adaptive response to injury. Uninjured cells elsewhere in the body react to these mediators by releasing compounds of their own, for example the release of adhesion molecules in the lung following haemorrhagic shock5. The result is a cascade of biochemical changes that collectively constitute a systemic disease. One result may be the dysfunction of organs that were not themselves damaged or ischaemic, such as the predictable development of acute lung injury after severe musculoskeletal trauma. The systemic inflammatory response syndrome that occurs after serious injury is recognized as a long-term consequence of shock, although in reality there are many more cells and mediators involved than just the immune system and known cytokines6.
Seminal to the recognition of haemorrhage as the trigger for a complex systemic disease is the immediate change in coagulation function that has been documented after severe injury. Numerous investigators, notably Brohi and colleagues7 in London and Hess and co-workers8 in Baltimore, have documented prolongation of prothrombin time associated with severity of injury, even before significant fluid resuscitation. Brohi et al.9 further noted that tissue injury and ischaemia can lead to increased fibrinolysis owing to activation of protein C. Recognition that coagulopathy is an early and predictable feature of severe injury, as well as a universal factor in hospital death from traumatic haemorrhage, has led to many of the recent changes in resuscitation strategy.
Other components in the pathophysiology of shock, such as perception of pain, lead to systemic inflammatory changes with the potential for long-term consequences3. Changes in nutrition, ambient temperature and even body position can play a role in the long-term pathophysiology of shock. Intruding into this already nearly incomprehensible cauldron of pathophysiology are the effects of resuscitation: the intended and unintended consequences of management for the injured patient.
Hazards of resuscitation
One of the basic lessons of traumatology is that no therapy comes without risk. Manipulation of broken bones and open wounds increases pain, changes area of local ischaemia and risks bacterial contamination. Tourniquets and pressure dressings work to stop blood loss, but may increase local tissue injury. Aggressive airway management risks the inability to intubate, exacerbation of tenuous haemodynamics, and even subtle dangers such as hyperventilation and reduced cerebral blood flow in patients with traumatic brain injury (TBI)10. In addition, fluid therapy is now recognized as a significant contributor to the overall pathophysiology of shock.
The hazards of administering fluids to severely injured patients fall into several categories, including the influences on temperature, volume in circulation, and effects of blood compounds such as clotting.
Most resuscitative fluids are given at ambient temperature; if given rapidly or in large quantities, the negative thermal load will overwhelm the patient's ability to compensate, producing significant metabolic strain at a time of great vulnerability. If compensation fails or is inhibited (by anaesthesia, for example) and hypothermia develops, a further cascade of physiological disruption occurs. One simple example of this is humoral coagulation. The primary rate constant for each individual clotting reaction is decreased by hypothermia, and the secondary synchronization of the whole cascade is disrupted. With an effect on every chemical reaction in the body, hypothermia is the quintessential pleiotrophic mediator, but it is only one unintended consequence of fluid resuscitation. Other negatives centre on the quantity of fluid administered and the qualities of the fluids used.
Bulk administration of intravenous fluid produces immediate mechanical effects on the circulation. In an injured patient who is intensely vasoconstricted, even as little as 100 ml fluid will increase end-diastolic pressure sufficiently to raise cardiac output and blood pressure11. Increased pressure may improve perfusion in non-constricted vascular beds, but may also reverse important haemostatic reflexes and exert excessive pressure on fragile extravascular clots12. Fluid administration leading to increased blood pressure in the setting of active haemorrhage predictably increases rebleeding and overall blood loss, a fact first noted during World War I and rediscovered episodically in every major conflict since13, 14.
As to fluid quality, it is obvious that administration of any intravenous product other than fresh homologous blood will change the composition of circulating blood, with potentially deleterious consequences. Isotonic crystalloids and various manufactured colloid solutions are the fluids traditionally available in the prehospital and emergency department (ED) environment. These dilute the concentration of red blood cells, clotting factors and platelets, contributing to coagulopathy. Bolus administration of crystalloids is associated with disruption of the endothelial glycocalyx, loss of membrane integrity, and cellular and extracellular oedema15. Reduced blood viscosity may improve perfusion, but also risks increased extravasation. In addition, the fluids themselves may have direct negative effects on the immune system, and on renal and hepatic function16.
Weighing against these concerns is the recognition that, in the absence of treatment, the severely injured patient will surely die. Resuscitation is required, but our understanding of how best to accomplish it, and how best to integrate surgical and non-surgical goals, has undergone near constant evolution in recent years. The following sections of this review present current recommendations for resuscitation of the severely injured patient, link each recommendation to the pathophysiology described above, and present the scientific evidence (or lack thereof) in support.
Overview of resuscitation during damage control surgery
Few would disagree that the more rapidly anatomical control of surgically accessible lesions is achieved, the better the outcome is likely to be. Adjuvant non-surgical therapies, collectively referred to as ‘damage control resuscitation’ or ‘haemostatic resuscitation’, are outlined in Table1.
Table 1. Elements of haemostatic resuscitation and subjective assessment of the supporting scientific literature
Deliberately maintaining a lower than normal blood pressure has been a mainstay of resuscitation strategy for almost two decades. An abundance of data from well constructed laboratory studies in a variety of small and large mammals has shown that survival from uncontrolled haemorrhagic shock is enhanced by fluid resuscitation limited to that required to maintain mean arterial pressure (MAP) at 50–60 mmHg (70–80 per cent of normal)17. Optimal oxygen delivery is achieved when some fluids are provided, but without attempting to normalize MAP. Animal models have been constructed that precisely delineate the boundaries of fatal hypoperfusion (about 2 h at a MAP of 40 mmHg) and fatal rebleeding (any MAP greater than 80 per cent of normal)19. Published studies almost universally demonstrate reduced total blood loss and improved survival when the haemorrhaging animal is maintained within these limits17, 19.
Limitations of the bench science literature include the use of blood pressure as a surrogate for tissue perfusion, and the less obvious effects of general anaesthesia required by the protocols used. Hypotension is much less likely to create critical tissue ischaemia when it occurs in the setting of vasodilatation (low pressure–high flow, as in the animal anaesthetized for a research study) than when it occurs in the setting of intense vasoconstriction (low pressure–low flow, as in an injured patient). This bias has perhaps led to a more favourable view of deliberate hypotension in the animal literature than might be justified clinically.
Two prospective randomized trials of deliberate hypotension have been conducted in human trauma victims, and another is under way at present. The first, led by Bickell in Houston in the early 1990s, randomized 598 hypotensive victims of penetrating torso trauma to fluid or no fluid during prehospital and ED care20. The no-fluid group required 2 litres less isotonic crystalloid, had a similar MAP at the time of admission, and achieved significantly improved survival to hospital discharge (70 versus 62 per cent; P = 0·04). The second trial, from the Shock Trauma Center in Baltimore, randomized 110 hypotensive patients in the trauma resuscitation unit to management at a systolic blood pressure of 80 versus more than 100 mmHg until definitive control of haemorrhage had been achieved21. This trial found a significant difference in fluid administered and MAP over the course of the study, but observed equivalent survival (92 per cent in both groups). The noteworthy 26 per cent difference in baseline survival between the two studies (overall survival in the Bickell study was 66 per cent) reflects the time of patient enrolment (field versus ED, allowing for better exclusion of moribund patients), as well as a decade of incremental improvement in trauma care, including the advent of DCS techniques. The study now under way in Houston uses a protocol similar to that of the Baltimore study (for example, titration of fluids to a lower than normal pressure), as well as modern theory regarding DCS and appropriate laboratory studies. Preliminary results from this trial have been released, and appear to favour deliberate hypotension22.
Collectively these trials—and the discussion they continue to provoke—have established the value of deliberate hypotension in trauma resuscitation, and this is the preferred approach in most major centres today. Patients who are actively bleeding, and in a damage control mode, should be maintained at a target systolic pressure of less than 100 mmHg18.
One unanswered question concerns the ideal target MAP in older patients and those with TBI. On the one hand, these patients are known to be at greater risk from periods of hypotension. On the other hand, they are also at greater risk from increased blood loss and longer periods of resuscitation. In a single canine study that examined resuscitation in a combined haemorrhage/TBI model, the best survival was achieved by early use of deliberate hypotension31. To date, no human data have been presented on this topic; the combination of severe TBI and active haemorrhage is known to carry a high fatality rate, but the heterogeneity of patients and injuries has so far prevented even an observational series to test the potential value of early deliberate hypotension. This point remains controversial.
Another important question for future study might be: What is the best way to achieve and sustain hypotension? Historically, this has been done by limiting fluid administration, allowing the patient to remain in a vasoconstricted state, and accepting a period of decreased perfusion in exchange for more rapid control of haemorrhage and less total blood loss. Understanding the physiology involved, however, has led to speculation that a better approach might be hypotension achieved through anaesthetic administration32. Induced vasodilatation allows for more liberal fluid administration and increased perfusion without raising the MAP, facilitating haemostasis without perpetuating tissue ischaemia. Although performed routinely in the author's centre over the past few years, this approach to deliberate hypotension has yet to be studied or validated in a prospective trial.
Management of blood content
Multiple theoretical threads have been spliced together in recent years to change the recommended approach to fluid choice during damage control resuscitation. One observation is that dilutional exacerbation of haemorrhagic shock can be minimized if resuscitation is performed with blood products rather than clear liquids. A second observation has been summarized in Fig.1, which shows the way in which a donated unit of whole blood changes through the process of fractionation and then theoretical reconstitution at the point of care. The resulting ‘whole blood’ fluid provides minimally acceptable levels of red blood cells (RBC), clotting factors and platelets, with scant margins to replace existing deficits. In other words, a resuscitation fluid that is composed completely of banked blood products is still relatively dilute. Third, the trauma and blood-banking community has awoken to the critical difference between Vietnam-era whole blood-based resuscitation (on which earlier literature and Advanced Trauma Life Support recommendations are based) and substitution of fractionated RBC: the absence of soluble clotting factors. Finally, the observation of coagulopathy as an important component of severe injury—one not seen in elective surgery—has prompted renewed interest in providing early support to the coagulation system33.
This led a decade ago to an evolution in thinking about fluid choice in actively bleeding patients in need of DCS: give no fluids that do not clot or carry oxygen. This approach is captured in the simple mnemonic ‘1 to 1 to 1’, and began to see anecdotal application in exsanguinating civilian patients with trauma in the early 2000s. The global War on Terror, beginning in 2002, focused further attention on this issue, and provided an unfortunate but valuable test-bed for empirical assessment of damage control techniques in general. The landmark paper by Borgman and colleagues23 in 2007 showed a dramatic difference in the mortality rate of injured soldiers requiring massive transfusion, based on the quantity of plasma that was co-administered. The mortality rate was 65 per cent in those receiving less than 1 unit plasma for every 4 units RBC, but only 20 per cent in those with a ratio of 1 : 2 or above.
The Borgman study was retrospective and observational, and conducted with the limited resources available in a war zone. A ‘survivor bias’ was clearly present, as soldiers bleeding more rapidly were likely to die after receiving RBC but before plasma could reach the bedside. Nonetheless, the differences in survival were dramatic, and similar observations were soon reported from civilian trauma centres and aggregated registries24. There have been more than 20 publications on this topic in the past 3 years25, 34, 35. Methodologies designed to account for early mortality show less impressive, but still positive, results26. Much debate has subsequently ensued: Which patients should be approached this way? What ratio of plasma to RBC is most effective? What diagnostic studies can be used as a guide?
Prospective trials are under way, but not yet available for interpretation. Clinical practice has changed, however, towards the earlier use of plasma (and platelets, although there is less literature on this topic) during DCS. The principles in Table2 represent the general consensus of informed traumatologists.
Table 2. Consensus statement for resuscitative strategies in trauma
Every trauma centre should have a massive transfusion protocol in place that reduces the logistical barriers to early administration of blood products26
Transfusion of patients who are not actively bleeding should be managed parsimoniously; blood ‘transplantation’ is a risk factor for systemic inflammatory dysregulation and subsequent organ system injury27
In patients with ongoing, life-threatening haemorrhage requiring damage control surgery, early transfusion of blood products may enhance haemostasis and minimize ischaemic injury
In bleeding injured patients, transfusion therapy should not be delayed while awaiting diagnostic results. Empirical use of uncross-matched ‘universal donor’ products is justified during early resuscitation and moments of diagnostic uncertainty28
The optimal ratio of plasma to red blood cells for empirical resuscitation during damage control surgery is controversial, but early support of coagulation is likely to be beneficial and may pre-empt the need for later, more massive, transfusions
A positive effect of this controversy has been increased communication between traumatologists and blood bankers. Good evidence and near-universal support exist for the concept of an institutional massive transfusion protocol to reduce the logistical barriers in delivery of blood products to the bedside of patients undergoing massive transfusion. This enables the clinicians to make more flexible and immediate choices about which products to administer and when. High-volume trauma centres have long extended this concept to allow for ready access to uncross-matched type O RBC for immediate use36; a few are now providing liquid type AB (universal donor) plasma as well27. This latter advance requires significant support from the blood bank and regional blood supplier, and careful daily maintenance to avoid wastage of a rare and precious resource. The goal of all of these efforts is to move transfusion decisions as close to the bedside as possible, and into the hands of those with the most accurate patient information and broadest trauma management experience.
Alternative procoagulant therapies
Empirical transfusion-based resuscitation can support the patient with ongoing haemorrhage, but may be inadequate to reverse existing coagulopathy arising from either severe injury and deep shock or intercurrent factors such as cirrhosis, congenital factor deficiency or use of anticoagulant medications. Recognition of the central role of coagulopathy in the pathology of haemorrhagic shock has led to a variety of alternative approaches to restore or preserve effective coagulation. These include use of topical haemostatic agents, antifibrinolytics and systemic clotting factor concentrates.
Topical haemostatic agents offer the potential for more rapid anatomical control of haemorrhage and would thus be inherently beneficial. Risks include indirect tissue injury (as from the exothermic reaction of zeolite with blood), vascular uptake of thrombin leading to embolic complications, and obscuring relevant anatomy in a way that makes future definitive repair more difficult37. Technical development of topical haemostatic compounds, dressings and cocktails is proceeding rapidly, and it is likely that high-benefit, low-risk products will soon be used routinely for the management of haemorrhage in the OR and ED, and in the field.
Antifibrinolytic therapy is based on the observation that fibrinolytic activity is increased after tissue injury and ischaemia7. The CRASH-2 (Clinical Randomisation of an Antifibrinolytic in Significant Haemorrhage) trial prospectively assessed the use of tranexamic acid versus placebo in 20 000 patients with haemorrhagic shock after trauma, and found a benefit in mortality when tranexamic acid was given within the first 3 h28. This may have been the result of better clot stability and reduced bleeding (although transfusion requirement and estimated blood loss were similar across groups), or of as yet undefined changes in inflammation and endothelial response to injury. The risks of antifibrinolytic therapy (thromboembolic complications) are relatively minimal, so this therapy appears justified early after injury in any patient at risk of prolonged haemorrhage.
Systemic clotting factor solutions, either single molecule (for example recombinant factor VIIa) or combinations (for example prothrombin complex concentrates), offer the advantage of supraconcentration over liquid plasma. Even a ‘mini-dose’ of factor VIIa (1 mg) is in the pharmacological range, providing 5–10 times the normal circulating concentration of the factor. Concentrates thus offer the possibility of reversing established coagulopathy more rapidly than can be accomplished with plasma alone. Case reports, observational series and a very few randomized trials have suggested various protocols for the use of these products to rescue patients undergoing DCS in whom more conventional therapy is failing29, 30. All have shown the potential rapidly to correct traumatic coagulopathy and reduce transfusion requirement in some patients and some situations, but none has yet been proven to reduce mortality. These products are unfortunately expensive, which has limited their availability and application, as well as the accumulation of knowledge regarding which patient and which moment is right.
The principal risk of factor concentrate therapy is the development of a thromboembolic complication. Severely injured patients who survive acute resuscitation have a high risk of these complications at baseline, which makes assessing the additional risk challenging. Thromboembolic complications after use of factor VIIa have been reported anecdotally and in at least one case series38. In this report, 9 per cent of injured patients exposed to the drug had a subsequent thromboembolic complication, of which one-third were temporally related. In the most recent prospective randomized trial of factor VIIa in injured patients, the overall incidence of thromboembolic complications was the same between groups, but there was an observable difference in the rate of arterial versus venous incidents, with more of the former in patients receiving factor VIIa39. Care is clearly necessary when administering these compounds to patients with known or strongly suspected arterial perturbations—patients with carotid or vertebral injuries, those with perivascular gunshot wounds, and those with active coronary artery disease.
Based on the lack of published data, it is difficult to make any blanket recommendation for the use of clotting factor concentrates during DCS. There are probably specific patients and situations when their use will be most helpful, and defining these populations and presentations is an ongoing focus of research. Given the heterogeneous presentation of injured patients and their injuries, the best possible recommendation is that the clinician at the bedside be familiar with these agents and how to obtain them, and consider their use when the benefit is likely to outweigh the risk.
Promoting endothelial integrity
A longstanding observation in patients who die from traumatic haemorrhage despite resuscitative efforts is that, in addition to the ‘fatal triad’ of coagulopathy, acidosis and hypothermia, the actual terminal moment is heralded by loss of reactivity in the vascular endothelium3. Inappropriate vasodilatation occurs and the patient's blood pressure becomes progressively unresponsive to either fluid boluses or administration of pressor agents. The experienced anaesthesiologist recognizes this as the moment when the patient's vital essence has slipped away, and further resuscitative efforts are likely to be futile. This observation, combined with the work of many current investigators on the mechanisms behind the coagulopathy of trauma, is increasingly highlighting the role of endothelial integrity in the pathophysiology of haemorrhagic shock.
Although humoral mechanisms are much better understood, the vascular endothelium plays a critical role in in vivo clotting. Upstream and downstream signalling in the vascular endothelium controls regional tissue blood flow, and thus both bleeding and perfusion. Endothelial cells are responsible for formation of the glycocalyx15 and for maintenance of cell–cell junctions that determine the degree of fluid extravasation out of the bloodstream. Signalling between the endothelium, platelets and RBC is critical to the initiation, rate and extent of clot formation at the local level, while the mediators involved probably have potent downstream and systemic effects as well40.
Understanding endothelial physiology, endothelial–humoral interactions, and the totality of the tissue response to injury and ischaemia is the bright frontier of trauma research in 2011. It is likely that some of the therapies we know to be beneficial but don't fully understand, such as administration of tranexamic acid, actually have their most important effect on the endothelium. It is also likely that future therapies designed or discovered to promote endothelial integrity, such as reduced sympathetic stimulation through general anaesthesia, will play an important role in recommendations for resuscitation during DCS in the decades to come.
Resuscitation strategies during DCS may be as important as the anatomical repair itself to long-term outcomes. Recommendations include avoidance of hypothermia, maintenance of a lower than normal blood pressure, and early support of the coagulation system in patients who are likely to require massive transfusion. Controversies include the optimal ratio of plasma to RBC for empirical resuscitation, the ideal role of clotting factor concentrates, and the potential benefit of early, deep anaesthesia on long-term outcomes.