resuscitated but in whom inotropic support is indicated. SPV and PPV parameters are surrogates for stroke volume variation (SVV). The magnitude of SVV broadly correlates to preload reserve [8]. Systolic pressure variation (SPV), which is the difference between the maximum and minimum systolic blood pressure during one mechanical ventilation cycle, has been demonstrated to reflect the degree of blood loss and the associated decrease in cardiac output (CO) during hemorrhage and to predict fluid responsiveness to volume loading. It should be stressed that the concept of PPV and SPV to assess preload reserve is only reliable in mechanically ventilated patients with a regular cardiac rhythm [9].

Urine output measured from a Foley catheter provides a valuable source of information during burn surgery but there are limitations that should be recognized. The American Burn Association guidelines state that urine output should be maintained at approximately 0.5–1.0 ml/kg/hr in adults and 1.0–1.5 ml/kg/hr in children [10] and urine flow rate is often used as an index of global perfusion during general anesthesia. However, oliguria ( > 0.5 mL/kg/ hr) as a predictor of hypovolemia or acute kidney injury is less reliable during general anesthesia than in non-anesthetized patients [11]. Anesthetic interventions, whether involving volatile agents, intravenous drugs, or regional blocks, may reduce blood pressure and cardiac output, leading to decreased glomerular filtration and urine formation. Urine output is not a reliable indicator of adequate intraoperative resuscitation in patients receiving diuretic therapy, inotropic support, high cardiac filling pressures, or abdominalcompartmentsyndrome.Incontrast,examination of urine provides a reliable indicator of hemolytic transfusion reaction. The signs and symptoms of a transfusion reaction are masked by general anesthesia or by the hemodynamic changes associated with burn surgery. Consequently, when an intraoperative transfusion is planned, a Foley Catheter should be used because the presence of hemoglobinuria may be the only reliable indicator of hemolytic transfusion reaction.

Monitoring of end tidal carbon dioxide (ETCO2) is an indicator of tracheal intubation and provides useful information regarding respiratory rate, airway resistance, and adequacy of pulmonary perfusion. In normal individuals, the difference between ETCO2

and PaCO2 is 2 to 5 mm HG. The gradient between end-tidal and arterial CO2 reflects dead space ventilation, which is increased in cases of decreased pulmonary blood flow, such as pulmonary air embolism or thromboembolism and decreased cardiac output. Therefore, ETCO2 monitoring can also provide important information regarding systemic perfusion, which may be perturbed in severely burned patients.

As hypothermia is a major concern, core temperature should be closely followed throughout the perioperative period. Cutaneous vasoconstriction is the major mechanism for heat retention and core temperature preservation in humans [12]. Burned patients who are undergoing surgical interventions are at high risk for hypothermia. Much of the cutaneous vasculature is severely damaged in patients suffering full thickness burns and, in many cases a large amount of skin has been excised as part of surgical treatment. These alterations ablate cutaneous mechanisms of heat conservation. In addition, general anesthesia causes major perturbations in central temperature control mechanisms. Specifically, it increases the central temperature set point that initiates heat conservation adaptations such as cutaneous vasoconstriction, brown fat metabolism and shivering. The adaptive responses to loss of central temperature are impaired in patients under general anesthesia. Therefore, it is important to maintain a warm operating room environment during burn surgery to minimize heat loss into the environment. Intravenous fluid warmers and airway warming devices can also be useful for minimizing heat loss but do not allow for active warming. Avoidance of hypothermia is an important goal in burned patients because hypothermia is associated with complications such coagulopathy, hypermetabolism and acute lung injury [13–14].

Pharmacology

The choice of anesthetic techniques and drugs is also dictated by the physiologic status of the patient. Large burn injuries are associated with profound physiological and metabolic changes that produce clinically significant alterations in responses to many drugs [15]. In some cases, drug doses must be reduced to avoid toxicity whereas doses of other drugs

must be increased or given more frequently to obtain a desired response. In the case of succinylcholine, large burn injuries lead to exaggerated, potentially lethal hyperkalemic responses and this drug is usually avoided in patients with acute burns. Both pharmacokinetic and pharmacodynamic variables are affected by burn injuries.

For one to two days after large burn injury, loss of intravascular fluids to the wound exudates and to the interstitium and, in some cases, myocardial depression lead to hypovolemia, decreased cardiac output, and increased vascular resistance [16]. These changes often result in decreased organ perfusion. As a result, drug responses may be enhanced and/or prolonged during this phase. The effects are offset to a variable degree by resuscitation efforts. Extreme volumes required for resuscitation result in severe edema and fluid shifts that alter drug volume of distribution and dilute drug binding plasma proteins. After two to three days, physiological status often changes dramatically with development of a hyperdynamic circulation and increased metabolic rate. Volumes of distribution may be increased and drug clearance may be augmented due to increased renal and hepatic perfusion (e. g. fentanyl and propofol) [17]. Burn injury causes opposite effects on the two major drug binding plasma proteins, albumin and alpha1-acid glycoprotein. Loss of albumin through the burn wound exudates and decreased hepatic synthesis result in decreased plasma albumin concentrations. In contrast, alpha1 – acid glycoprotein is one of the acute phase proteins and its concentration can double after major burns. Albumin binds more acidic drugs (e. g. diazepam or thiopental) while alpha1-acid glycoprotein binds mostly basic drugs (e. g. propranolol, lidocaine, or imipramine). As result, protein binding of drugs bound mainly to albumin is reduced while protein binding can be increased for drugs bound mainly to alpha1-acid glycoprotein. Large open wound surfaces provide a novel route of drug elimination and clinically significant elimination of drugs in the wound exudate has been reported [18].

Among alterations in responses to anesthetic drugs used in burn patients, clinically significant alterations in responses to muscle relaxants are probably the most commonly encountered. Large burn injuries are associated with increased expression of

nicotinic acetylcholine receptors across the surface of skeletal muscle. This leads to decreased sensitivity to non-depolarizing muscle relaxants requiring higher and more frequent doses [19]. An exception is Mivacurium, for which the pharmacodynamic change in nicotinic receptors on the skeletal muscle surface is offset by a pharmacokinetic change due to decreased plasma cholinesterase that prolongs the action of Mivacurium in burn patients and, as a result, dose requirements are the same as for nonburned patients [20].

In contrast to the non-depolarizing relaxants, large burns cause sensitization to succinylcholine and an exaggeration of its hyperkalemic effect. This hyperkalemic effect has been associated with cardiac arrest in some patients and, as a result succinylcholine is generally avoided in burn patients. The question of when succinylcholine can be safely administered after burn injury is controversial. Since there is little clinical experience with succinylcholine administration earlier than two weeks after burn injury, recommendations are made largely on extrapolation of observations from experimental studies. Authorities have recommended that succinylcholine is safe to administer, from the standpoint of exaggerated hyperkalemia, for approximately 48 hours after large burns [21]. Since resistance to non-depolarizing relaxants have been reported up to a year after injury, it is also recommended that succinylcholine be avoided for a year after wounds have healed.

All these changes, along with the dynamic nature of their development and resolution over time, make it very difficult to predict alterations in drug responses with any degree of precision. The key is to carefully monitor drug responses (and drug level when necessary) and titrate doses according to patient responses.

Perioperative fluid management

Fluid management during debridement and grafting of burn wounds is one of the more challenging aspects of perioperative care of burn patients. Perioperative care involves not only replacement of shed blood during surgery but the ongoing ICU fluid management of a critically ill patient. If the patient is

taken to the operating room early after the injury, practitioners must consider the resuscitation needs of the patient during the burn shock phase. Burn shock results from the transudation of plasma from the intravascular space into the interstitial space in response to massive cutaneous injury and inflammation and results in significant intravascular hypovolemia. Days and weeks after the injury, perioperative fluid management must be executed with regard to the patient’s hospital course and recent fluid management choices in the burn ICU. Perioperative care should be coordinated with ICU management. In the days following initial resuscitation, diuresis and fluid restriction are commonly initiated to mobilize fluids and reduce edema. In this context, it is not helpful when a large amount of crystalloid solution is administered during burn wound excision. In addition, surgeons often utilize a tumescent technique to facilitate wound debridement and skin graft harvest by injecting a dilute crystalloid solution subcutaneously (i. e. clysis solution) [22]. The volume of this solution can be quite large and should be considered a source of perioperative fluid administration. In general, limiting the amount of crystalloid solution given to support cardiac preload will help minimize edema and the need for post-operative diuresis.

When extensive burn wounds are excised, blood loss can be brisk and involve substantial volumes. Volume replacement is guided by at least three separate goals: maintain cardiac preload to support cardiac output, administer red blood cells to provide oxygen carrying capacity, and, in the case of massive hemorrhage, administer fresh frozen plasma (FFP) or other components to replace coagulation factors that are lost due to dilution and consumption.

Titrating fluids for volume replacement during burn surgery is difficult for several reasons. It is not possible to accurately estimate the volume of ongoing blood loss during major burn wound excision. Shed blood is concealed beneath the patient or within dressings and may be spread over a broad surface. The patient’s physiological status can change rapidly as blood volume is lost or as inflamed and infected wounds are manipulated and release microorganisms and inflammatory mediators with hemodynamic effects that alter myocardial contractility and/or vascular compliance. There is no single physi-

ological variable that can be consistently relied on for titration of administered fluids. The anesthetist must continuously monitor filling pressures, blood pressure, arterial wave form (when measuring arterial blood pressure directly), urine output, hematocrit, and blood gas analysis while replacing shed blood during burn wound excision.

Normal saline (0.9%) has been a popular solution for intravenous volume replacement. Large amounts of saline administered intravenously, however, have been associated with hyperchloremic acidosis. This otherwise relatively benign condition can confuse assessment and/or exacerbate effects of acidosis due to poor tissue perfusion [23]. Use of lactated Ringer’s solution avoids this problem but there is a theoretical risk of formation of micro thrombi when lactated Ringer’s solution is used to dilute packed red blood cells for transfusion.

A variety of colloid solutions are available for volume replacement. Albumin is widely used but is occasionally in short supply and, in the past, cost concerns have limited its use. Hydroxyethyl starch preparations are frequently preferred. These hetastarch solutions are available in a variety of preparations differing in concentration, molecular size, molecular substitution, and position of the hydroxyethyl group (C2 vs. C6 position on the glucose molecule) [24]. Newer preparations with smaller molecular weight and less molecular substitution may have fewer undesirable effects such as impaired coagulation. Judicious use of colloid solutions can help minimize the volume of crystalloid administered, limiting edema formation and facilitating postoperative care in the burn ICU.

The transfusion trigger for administering red blood cells varies considerably between patients and there is no general agreement regarding indications for transfusion of burn patients. In the past, hemoglobin concentration has been maintained at 10g/100 ml or above for patients with large burns. More recently, lower hemoglobin concentration has been accepted to reduce the exposure of patients to allogenic blood and to preserve the blood bank resources. For patients with small burns and without co-existing disease processes, 6–6.5g/100 ml may be tolerated while patients with co-existing cardiac or pulmonary disease may require 10g/100 ml. If the starting hematocrit and tissue perfusion appear ad-

equate, initial volume replacement can be accomplished with colloid and this will result in a decrease in the hematocrit. Shed blood then contains a smaller red blood cell mass and less blood will need to be transfused to return the hematocrit to the desired level. Periodic blood gas analyses can be used to monitor the patient’s tolerance of the reduced oxygen carrying capacity and help guide decisions of when to transfuse packed red blood cells.

An often overlooked role of red blood cells is their ability to facilitate hemostasis. Red blood cells have a rheological effect that increases the margination of platelets by pushing them to the periphery of the blood vessel increasing the near wall concentration of platelets and enhancing their interaction with injured endothelium. Red blood cells may also have effects on platelet biochemistry and responsiveness to enhance their role in hemostasis [25].

Coagulopathy is one of the more consistently observed complications of massive hemorrhage and transfusion. Volume replacement with fluids lacking coagulation factors results in dilutional coagulopathy. Guidelines of the American Society of Anesthesiologists (ASA) Task Force on Perioperative Blood Transfusion and Adjuvant Therapies recommend administration of FFP to patients with micro vascular bleeding when the PT, INR or aPTT are elevated. The guidelines state that FFP is not indicated to treat bleeding when these laboratory measurements are normal [26]. However, recent clinical studies suggest that traditional guidelines for administration of fresh FFP may be suboptimal for managing coagulopathy associated with massive hemorrhage and transfusion. Generally accepted definitions of massive blood loss are one blood volume in 24 hours, 50% blood volume in 3 hours, or ongoing blood loss of 150 ml/min. At these rates of blood loss and volume replacement, coagulation factors are rapidly diluted and hemostasis may be impaired. Hirshberg et al. (2003) have used a computer simulation of exsanguinating hemorrhage to estimate changes in prothrombin time, fibrinogen, and platelets and the efficacy of various resuscitation strategies in supporting coagulation function [27]. Their simulation indicated that current protocols for massive transfusion do not provide adequate coagulation factor replacement to prevent or correct dilutional coagulopathy. They found that in order to prevent

coagulopathy in their model, it was necessary to give plasma before the prothrombin time was elevated and the optimal ratio of FFP to packed red blood cells was 2:3. Subsequent clinical studies in both military and civilian trauma patients have found significant dose-dependent decreases in mortality when plasma was administered early and the ratio of FFP to packed red blood cells was increased. Clinical experience has shown that once coagulopathy develops it may be difficult to reverse and the presence of coagulopathy has been associated with increased morbidity. ASA guidelines for administration of FFP require the diagnosis of coagulopathy before FFP is administered. To prevent, rather then treat, coagulopathy in the massively hemorrhaging patient, Hirshberg et al. recommend a more anticipatory definition of massive transfusion such as transfusion of 4 units of packed red blood cells within an hour with anticipation of ongoing blood loss. A problem with the use of laboratory measurements of coagulation function during massive transfusion is that the results cannot be provided in a timely fashion. Under these circumstances, indications for administration of FFP cannot depend on lab measurements and decisions must be made empirically using protocols. As an example, in our institution, blood loss is replaced with colloid solution to support cardiac preload and packed red cells to maintain oxygen carrying capacity until 50% of the blood volume has been replaced. From that point forward, FFP and packed red blood cells are administered in a 1:1 ratio to replace ongoing blood loss. One key in these decisions is the anticipation of ongoing blood loss, which is determined by clinical judgment.

In addition to dilutional coagulopathy, there are several other clinical problems associated with massive blood transfusion. Hypothermia is a risk during burn surgery and this can be exacerbated by rapid administration of fluids that are inadequately warmed. It is important that blood warmers capable of warming fluids at the flow rates required for resuscitation of massive hemorrhage are used (e. g. >50 ml/min in adults). Rapid infusion of large volumes of blood products, especially FFP, can cause significant ionized hypocalcaemia due to citrate toxicity. Reduced ionized calcium levels affect vascular tone, cardiac contractility, and coagulation. This is more rapidly treated with calcium chloride rather than calcium gluconate