Low-volume fluid resuscitation for presumed hemorrhagic shock: helpful or harmful?

Susan A. Stern, MD

For the past 4 decades, the standard approach to the trauma victim who is hypotensive from presumed hemorrhage has been to infuse large volumes of fluids as early and as rapidly as possible. The goals of this treatment strategy are rapid restoration of intravascular volume and vital signs towards normal, and maintenance of vital organ perfusion. The most recent laboratory studies and the only clinical trial evaluating the efficacy of these guidelines however, suggest that in the setting of uncontrolled hemorrhage, todays practice of aggressive fluid resuscitation may be harmful, resulting in increased hemorrhage volume and subsequently greater mortality. This has been demonstrated in animal models representative of penetrating trauma as well as those representative of blunt trauma. The data strongly suggest that limited or hypotensive resuscitation may be preferable for the trauma victim with the potential for ongoing uncontrolled hemorrhage. Limited resuscitation provides a mechanism of avoiding the detrimental effects associated with early aggressive resuscitation, while maintaining a level of tissue perfusion that although decreased from the normal physiologic range is adequate for short periods. Large randomized clinical trials are necessary to confirm this new laboratory data. Future research should focus on developing resuscitation methods that may actually enhance tissue perfusion during limited resuscitation and therefore offset its potential detrimental effects. Curr Opin Crit Care 2001, 7:422430 2001 Lippincott Williams &
Wilkins, Inc.

For the past 4 decades, the standard approach to the trauma victim who presents hypotensive from presumed hemorrhage has focused on early aggressive resuscitation with large volumes of crystalloid, and blood products as deemed appropriate [1]. The goal of this treatment is to restore intravascular volume and vital signs back to normal as quickly as possible to maintain vital organ perfusion. This has been the approach regardless of whether the victim is bleeding from a readily controllable source such as an extremity, or from an inaccessible injury to the chest or abdomen. The rationale for these recommendations are based on interpretation of data from experimental work of the 1950s and 1960s [27]. Investigators were interested in developing a hemorrhage model that resulted in irreversible shock, defined as a physiologic state following hemorrhage in which the animals would die despite aggressive resuscitation. To accomplish this, animals were bled from an intravascular catheter and maintained at a prescribed level of hypotension for varying but always prolonged periods of time prior to initiating resuscitation. This model, known as a Wiggers preparation, became the standard for the study of hemorrhagic shock and resuscitation. Investigators made several observations during these experiments. First, they observed that to induce a state of irreversible shock required that animals be subjected to several hours of hypotension which included at least a brief period (i.e., 45 minutes) of severe hypotension, defined as a mean arterial pressure (MAP) of 3040 mmHg [2]. Second, they observed that prolonged hemorrhagic hypotension resulted in a marked extracellular fluid (ECF) deficit, that this ECF deficit was corrected only by the administration of isotonic crystalloid in volumes of 23 times the estimated blood loss, and that correction of this deficit greatly improved survival [37]. This is the basis for the well-known 3:1 rule and the current emphasis on volume restoration for the treatment of acute hemorrhagic hypotension. Recently investigators have begun questioning the current recommendations for aggressive volume replacement in the trauma victim with ongoing uncontrolled hemorrhage. First, they note the complete lack of randomized clinical trials supporting this practice. Second, they suggest that while previous laboratory studies have helped greatly in our understanding of several aspects of the physiology of shock, the models they used are not representative of the trauma patients encountered in the

Department of Emergency Medicine, University of Michigan, Ann Arbor, Michigan, USA. Correspondence to Susan A. Stern, MD, Department of Emergency Medicine, University of Michigan, TC-B1354; Box 0303, 1500 East Medical Center Drive, Ann Arbor, MI 481090303; e-mail: suestern@umich.edu Current Opinion in Critical Care 2001, 7:422430 Abbreviations MAP ECF CNS RL SVRL LVRL PP PRBCs HBOCs DCLHb mean arterial pressure extracellular fluid central nervous system Ringers lactate small volume resuscitation with Ringers lactate large volume resuscitation with Ringers lactate pulse pressure packed red blood cells hemoglobin-based oxygen carriers diaspirin cross-linked hemoglobin

ISSN 10705295 2001 Lippincott Williams & Wilkins, Inc.

422

Low-volume fluid resuscitation Stern 423

pre hospital setting or emergency departments today. The two most common causes of death among civilian and battlefield casualties are rapid exsanguination and CNS injury, not protracted hypotension [813]. And third, they note that the Wiggers model of hemorrhage used in these studies does not adequately reproduce the pathophysiology of the acutely exsanguinating trauma victim. In a Wiggers preparation blood pressure is controlled by the investigator rather than the animals response to the hemorrhagic insult, i.e., blood pressure is an independent variable. In addition, in this type of model animals are hemorrhaged solely from an intravascular catheter; bleeding is stopped at the will of the investigator via a simple turn of a stopcock. Hence these have been termed controlled hemorrhage models. In the clinical setting however, hemorrhage occurs because of a disruption in the vascular circuit, and the potential for ongoing uncontrolled hemorrhage remains until surgical hemostasis is obtained.

strate exacerbation of hemorrhage if aggressive fluid resuscitation is attempted prior to obtaining surgical hemostasis [14,15,17,19,20]. It is appropriate to note, however, that data from studies which use models of isolated arterial or venous injury may not be applicable to the blunt trauma victim. In an attempt to better represent the latter patients, investigators have developed solid organ injury models of uncontrolled hemorrhage. In a model of moderate splenic injury in rats, untreated controls sustained significantly less hemorrhage volume as compared with animals aggressively resuscitated with RL [23]. Solomonov et al. developed a more severe model of splenic injury in rats, and similarly demonstrated that animals aggressively resuscitated with isotonic crystalloid experienced greater hemorrhage volumes and significantly lower survival times as compared with untreated controls [22]. Matsuoka et al. used a model of liver injury in rats, in which uncontrolled hemorrhage is created by sharp resection of 65% of the median and left lateral lobes of the liver, to compare the effects of small (SVRL) versus large (LVRL) volume resuscitation with RL [24]. Intraperitoneal hemorrhage volumes were significantly greater in the LVRL as compared with both SVRL treated animals and untreated controls. Four-hour mortality rates were 50%, 47%, and 40% for controls, SVRL, and LVRL treated animals, respectively (P > 0.05). Limited conclusions can be drawn from this mortality data however, as the control and SVRL groups received no additional fluids for the entire 4-hour period. Of note is that mean survival times for both controls and SVRL treated animals was approximately 3 hours, and would have been more than adequate for most trauma victims to reach a facility capable of providing definitive surgical care. In all three of these models, the hemodynamic response to aggressive resuscitation mirrored that observed in vascular injury models [2224]. Blood pressure increased abruptly following initiation of resuscitation, but then rapidly declined to levels equal to or less than that of controls. Therefore, similar to models of isolated arterial or venous injury, aggressive resuscitation with isotonic crystalloid in a model representative of blunt trauma, resulted in significantly greater hemorrhage volume and no improvement in mortality as compared with limiting or withholding resuscitation.

Current guidelines may be harmful

To address the limitations of previous laboratory studies, investigators developed uncontrolled or vascular injury hemorrhage models. In an uncontrolled or vascular injury hemorrhage model, animals are bled either through a fixed vascular lesion or via a standardized solid organ injury, i.e., the spleen or liver. Hemorrhage volume and duration depend on the animals physiologic responses, i.e., thrombus formation and vasoconstriction, rather than on the simple closure of a stopcock. Because there is potential for ongoing or recurrent hemorrhage, the investigator can evaluate the effect of the intervention on the injury site and on the organisms natural hemostatic mechanisms. Several uncontrolled hemorrhage models have been developed, and regardless of the model, almost all of these studies demonstrate results opposite the classic controlled hemorrhage studies; that is they suggest todays practice of aggressive fluid resuscitation may be harmful, resulting in increased hemorrhage volume and subsequently greater short-term mortality [1418,20,22 24,26,27,30,74] (Table 1). Bickell et al. were the first to evaluate the effects of conventional rapid volume expansion in a large animal vascular injury model of hemorrhage, in which swine were bled from a 5 mm tear in the infrarenal aorta [18]. Control animals, those not resuscitated, experienced significantly lower hemorrhage volumes (783 85 and 2142 178 mL), and 2-hour mortality (0% and 100%) as compared with those resuscitated with Ringers lactate (RL) in a volume equal to three times the estimated blood loss. This data actually suggest that one can make a non-lethal injury, lethal by over-resuscitating. Several other studies, utilizing models of either arterial or venous injury, support the data from Bickell et al. and demon-

Limited resuscitation is an appropriate option

Based on these studies, several investigators have suggested that resuscitation be withheld completely until operative control of hemorrhage is achieved. This may be a reasonable approach for the trauma victim who has experienced only a mild or moderate hemorrhage (as in Bickells model in which hemorrhage volume prior to resuscitation was approx. 25 mL/kg) [18], but what about the patient who has suffered a severe hemorrhage and is

424 Injury model Tail resection Incision of ileocolic artery Tail resection Laceration of lumbar veins Aortic puncture Aortic tear Aortic tear Aortic tear Aortic tear Laceration of pulmonary vein Liver resection Moderate splenic injury Severe splenic injury Rat Rat Rat SVRL vs LVRL vs HTS vs NR NR vs LVRL NS vs HTS vs NR Pig Sheep Bolus vs slow infusion of HSD RL vs NR Pig Pig Pig RL vs NR RL vs HSD vs NR NS to MAP = 40, 60, 80mmHg Rat Rat RL10 vs RL30 vs NR NR vs RL at varied rates and times Rat Rat Rat NS vs HTS vs NR NS vs HTS vs NR NS40 vs NS80 vs NR HV w/HTS HV w/HTS HV w/NS80 vs NS40 and NR HV w/RL10 and RL30 vs NR HV w/early & rapid RL infusion HV w/RL HV w/RL and HSD HV w/resusc to MAP = 80mmHg HV w/bolus infusion HV w/RL HV w/HTS and LVRL HV with LVRL vs NR HV with NS vs NR Animal Resuscitation regimens tested Effect on hemorrhage volume Effect on mortality or survival time Mortality w/HTS Mortality w/HTS 6-hour Mortality w/NR vs NS40 and NS80 No effect Mortality w/early rapid RL infusion Mortality w/RL Mortality w/RL and HSD Mortality w/resusc to MAP = 80mmHg Mortality w/bolus infusion No effect Mortality w/HTS No effect Surv Time w/NS and HTS

Table 1. Studies of uncontrolled hemorrhage

Study

Small Animal Vascular Injury Models Gross et al. [14]* Gross et al. [15]* Sindlinger et al. [16]

Smail et al. [17] Leppaniemi et al. [30] Large Animal Vascular Injury Models Bickell et al. [18] Bickell et al. [74] Stern et al. [26]

Stern et al. [27] Sakles et al. [20] Solid Organ Injury Models Matsuoka et al. [24] Krausz et al. [23]** Solomonov et al. [22]

*NS Infusion Volume = 5 mL/kg; HTS Infusion Volume = 5 mL/kg NS40 = Infusion volume of 40 mL/kg; NS80 = Infusion volume of 80 mL/kg; Note: In this study, the NR group received no fluids for 6 hours. NR animals had no mortality until after 60 minutes. RL10 = Infusion volume of 10 mL; RL30 = Infusion volume of 30 mL; Note: Although there was no difference in mortality, total hepatic blood flow was greater with RL10 vs. RL30. Slow Infusion = 0.4 mL/kg/min; Rapid Infusion = 1.33 mL/kg/min SVRL Infusion = 4 mL/kg; LVRL Infusion = 24 mL/kg; HTS Infusion = 4 mL/kg **LVRL Infusion = 41.5 mL/kg HSD, 7.5% NaCl/6% Dextran-70; HTS, 7.5% NaCl; HV, hemorrhage volume; NR, no resuscitation; NS, normal saline.

Low-volume fluid resuscitation Stern 425

near circulatory collapse? Completely withholding resuscitation from these patients may not be appropriate, and in all likelihood would result in death. And in fact, there are data that suggest that completely withholding resuscitation may be equally as harmful as overly aggressive resuscitation, and unnecessary. Kowalenko et al. compared the effects of low (MAP = 40 mmHg) and normal (MAP = 80 mmHg) pressure resuscitation in a near-fatal uncontrolled hemorrhage model [19]. Pigs with 4-mm aortic tears were bled to a pulse pressure (PP) of 5 mmHg and then resuscitated to a target MAP of 40 or 80 mmHg. Control animals were not resuscitated. Total hemorrhage volumes were 47 8, 45 12, and 78 28 mL/kg, while 1-hour mortality rates were 87.5%, 12.5%, and 62.5%, for Controls, those subjected to low pressure (MAP = 40 mmHg) and normal pressure (MAP = 80 mmHg) resuscitation respectively. Data from Capone et al., in which uncontrolled hemorrhage was created by 75% tail resection in rats, are consistent with the study by Kowalenko et al. [25]. While these data do demonstrate that aggressive fluid resuscitation in the setting of uncontrolled hemorrhage is associated with a poor outcome, the data also suggest that for a severe hemorrhagic insult, withholding resuscitation can similarly be fatal, and that limited or hypotensive resuscitation may be a more ideal approach. Stern et al. used a near-lethal swine aortic tear model to compare the effects of incremental increases in blood pressure during resuscitation of uncontrolled hemorrhage [26]. Following hemorrhage, animals were resuscitated to a MAP of 40, 60, or 80 mmHg. Total hemorrhage volumes and 1-hour mortality rates were significantly greater in animals in which normotension (MAP = 80 mmHg) was the therapeutic endpoint, but were not different between animals resuscitated to a target MAP of 40 or 60 mmHg. Animals subjected to severe under-resuscitation (MAP = 40 mmHg) however, were significantly more acidotic as compared with moderately under resuscitated (MAP = 60 mmHg) animals. Serum bicarbonate levels decreased to 8.2 3.7 and 15.5 1.5 mmol/L in animals resuscitated to a MAP of 40 and 60 mmHg respectively (P < 0.05). These data suggest that severe under-resuscitation, and the associated drastic tissue perfusion deficit, may not be necessary to avoid the detrimental effects of aggressive resuscitation. Moderate under resuscitation only minimally increased hemorrhage volume and resulted in significantly improved markers of tissue perfusion. Data from other studies suggest that alternative resuscitation strategies might allow one to avoid the detrimental effects of early aggressive fluid infusion while maintaining tissue perfusion. One such method is slow infusion of a hypertonic solution. Stern et al. used a near-lethal swine

aortic tear model to compare bolus to slow infusion of 8 mL/kg of HSD [27]. Hemorrhage volumes were 79 11 and 43 9 ml/kg (P < 0.001), while 90-minute mortality rates were 75% and 12.5% (P = 0.04) for the bolus and slow infusion treated animals, respectively. Aortic blood flow and MAP increased sharply immediately following initiation of resuscitation in the bolus treated animals and were significantly greater as compared with the slow infusion group during the first 15 minutes of resuscitation, but then rapidly declined. In contrast, slow infusion resulted in gradual increases in MAP, aortic blood flow, cardiac indices, and O2 delivery that eventually surpassed those in the bolus treated animals and approached baseline levels by 30 to 45 minutes. Burris et al. made similar observations in a study of rats subjected to aortic injury and resuscitated with 7.3% NaCl / 6% hetastarch (HH) to a target MAP of either 40 mmHg (HH40) or 80 mmHg(HH80) [28]. Seventy-two hour survival rates for controls, HH40 and HH80 treated animals were 0%, 67%, and 27% respectively. The HH40 treated animals experienced a gradual rise in blood pressure, ultimately overshooting the target MAP of 40 mmHg and reaching levels as high as the HH80 animals but at a significantly later time point and with a trend toward a lower recurrent bleed volume. These data suggest that greater increases in blood flow and pressure may be tolerated without accentuating hemorrhage if achieved gradually and at a time significantly delayed from initial injury. This concept is also supported by animal studies by Krausz et al. [29] and Leppaniemi et al. [30] which demonstrated that earlier initiation of resuscitation and faster infusion rates of the same resuscitation agent were associated with significantly greater hemorrhage volumes and mortality. That resuscitation strategies which cause abrupt increases in blood pressure and flow may increase hemorrhage volume is consistent with what is known about the pathophysiology of hemostasis. Although thrombus is formed almost immediately following an arterial injury, it is initially soft and jelly-like and inadequate to maintain hemostasis in high-pressure vessels. Transformation to a more rigid hemostatic plug requires significant fibrin deposition, which does not occur until at least 2030 minutes following injury [3136].

Clinical trial of delayed versus immediate resuscitation

Although there has been steadfast adherence to the current resuscitation guidelines for several decades, the only randomized controlled clinical trial evaluating the efficacy of these recommendations does not support them [37]. This prospective clinical trial compared immediate standard resuscitation to delayed resuscitation in 598 adults with penetrating torso injuries and a pre hospital blood pressure 90 mmHg. Delaying fluid resuscitation until the time of operative intervention resulted in improved survival to hospital discharge (70% vs. 62%) and

426 Trauma

a decrease in the length of hospital stay. This study had several limitations, however. First, the causes of death were not well defined. Second, there were several protocol violations; twenty-two (8%) patients in the delayed resuscitation group received fluid infusions prior to operative intervention in violation of the protocol. Finally, although all patients in this study had sustained penetrating torso injury, the severity of the shock state varied greatly. Systolic blood pressure ranged from being barely palpable to 90 mmHg. As suggested by previous data, the trauma victim in near circulatory collapse represents a very different physiologic entity from the victim with only mild to moderate hemorrhage, and the appropriate resuscitation regimen is also likely to be very different. Despite these limitations, it is significant that most patients in whom resuscitation was withheld for greater than 1 hour, survived to hospital discharge without major complications.

Advocates of early aggressive resuscitation of hypotensive trauma victims argue that the needs for optimizing vital organ perfusion exceeds any risk of accentuating hemorrhage. This argument is based on the supposition that hypotension from acute hemorrhage represents a state of shock which will certainly progress to endorgan failure and ultimately death if left untreated for even a brief period of time. The most recent laboratory data along with the clinical trial by Bickell et al. suggest that this assumption is not correct, and that substantial periods of moderate hypotension can be tolerated without resulting in significant tissue injury or end-organ failure.

The future: methods of improving outcome from limited resuscitation

While there is abundant data to suggest that hypotensive or limited resuscitation may be preferable to the current standard of care, a more optimal strategy would be one that actually enhanced tissue perfusion while minimizing ongoing hemorrhage. The latter might be accomplished through limited resuscitation using an agent with oxygen carrying capacity. An obvious choice would be whole blood or packed red blood cells (PRBCs). And in fact there is some evidence to suggest that early transfusion obviates some of the detrimental effects of limited resuscitation. In a study of swine undergoing hypotensive resuscitation following near-fatal uncontrolled hemorrhage, the administration of blood as compared with 0.9% NaCl as the initial resuscitation agent significantly reduced the severity of metabolic acidosis [40]. Similarly, in a study of uncontrolled hemorrhage and resuscitation in rats, hypotensive resuscitation with a combination of blood and crystalloid as compared with crystalloid alone resulted in a trend toward decreased lactic acidosis [41]. Unfortunately, blood products are not readily available in the pre hospital or battlefield settings. Potentially promising substitutes for blood replacement and enhancing cellular perfusion in the pre hospital and battlefield settings are the hemoglobin-based oxygen carriers (HBOCs). Data from two recent laboratory trials suggest that the HBOCs may be effective adjuncts to limited resuscitation from hemorrhagic shock. McNeil et al. used a controlled hemorrhage model to compare the efficacy of hypotensive resuscitation with the bovine HBOC product, HBOC-201 (Biopure Corporation, Cambridge, MA), to hypotensive resuscitation with blood [42]. Five-hour survival rates were similar between groups as were serum lactate and base deficits. Manning et al. used a model of uncontrolled exsanguinating liver injury in swine to compare the effects of low pressure resuscitation (target MAP = 60 mmHg) with HBOC-201 to low pressure resuscitation with RL solution [43]. Total hemorrhage volumes were not different between groups, however, two-hour survival rates were significantly

Long-term studies of limited resuscitation

Despite the increasing volume of laboratory data and the clinical trial by Bickell et al. which support a less aggressive approach to resuscitation of the trauma victim with uncontrolled hemorrhage, several questions still remain, not the least of which is what will be the long-term effects of limited or hypotensive resuscitation? What is the likelihood that this approach may improve short-term survival, but days later will be associated with the development of multisystem organ failure or neurologic compromise? Carrillo et al. evaluated the effects of hypotensive resuscitation on neurologic outcome in rats [38]. Animals were subjected to controlled hemorrhage and maintained at a MAP of 40 mmHg for 60 minutes, or 30 mmHg for 45 minutes and then resuscitated and allowed to recover. Ten-day survival rates were 91%, 59%, and 71% for Shams, and animals maintained at a MAP = 40 mmHg for 60 minutes, and for those maintained at a MAP = 30 mmHg for 45 minutes, respectively. All surviving animals tested normal neurologically and were without histologic evidence of neuronal loss. Stern et al. evaluated the long-term effects of moderate and severe under resuscitation on end-organ injury in a swine aortic tear model [39]. Following hemorrhage, animals were subjected to a 75-minute period of severe (target MAP = 40 mmHg) or moderate (target MAP = 60 mmHg) under resuscitation. Control animals were not resuscitated. After 75 minutes, the aortic injury was repaired and all animals were aggressively resuscitated. Seventy-two hour mortality rates were 100%, 36%, and 22% for controls, and animals severely and moderately under resuscitated, respectively. Surviving animals demonstrated no clinical or histologic evidence of endorgan injury.

Low-volume fluid resuscitation Stern 427

greater in the HBOC-201 treated animals (100% vs. 10%). The HBOC-201 treated animals also experienced less severe metabolic acidosis as demonstrated by lower serum lactate levels. These data suggest that the HBOCs may provide a mechanism of enhancing tissue perfusion and reducing oxygen debt, and therefore improve outcome from limited resuscitation of uncontrolled hemorrhage. Although animal studies with various stroma-free hemoglobin solutions appear promising, data from human trials using these same solutions have not been as encouraging. Several preclinical investigations suggested diaspirin cross-linked hemoglobin (DCLHb) might be efficacious in the resuscitation of the acute trauma patient [4248]. Unfortunately, in contrast to those promising studies, a multicenter randomized controlled trial assessing the effect of DCLHb in patients with severe traumatic hemorrhagic shock was stopped after interim analyses revealed a significant increase in mortality in the DCLHb treated patients as compared with the control group (Forty-eight hour mortality: 38% versus 15%, P = 0.01; Twenty-eight day mortality: 46% versus 17%, P = 0.003) [49]. An alternative to enhancing tissue oxygen delivery would be any therapeutic modality that could protect the organism against the ischemic effects of hypotensive resuscitation. Therapeutic hypothermia might afford this protection. Hypothermia has been shown to protect and preserve the brain and other vital tissues during global ischemic insults such as cardiac arrest, has shown promise in the treatment of traumatic brain injury, and is used for cerebral protection in certain surgical procedures [50,51]. Suggested mechanisms for the beneficial effects of therapeutic hypothermia include reduction of oxygen demand, adenosine triphosphate loss, leukotriene production, edema, free radical reactions, and capillary permeability [5158]. Despite evidence that hypothermia may protect vital organs during ischemia, concerns of potential detrimental effects have created a reluctance to consider its use in patients with traumatic hemorrhagic shock. These concerns are based largely on data from retrospective studies in which hypothermia developed accidentally in association with trauma and hemorrhage [5962]. These studies demonstrate a correlation between the development of spontaneous hypothermia and mortality. However, the data do not discern the true effects of hypothermia on outcome from other confounding factors, such as exposure, tissue trauma, large volume resuscitation, and anesthetic use. Despite this negative data, there has been renewed interest in the use of therapeutic hypothermia for the resuscitation of patients with hemorrhagic shock, and data from several animal studies suggest that this treatment modality may be beneficial [6369].

Kim and colleagues, in a rat model of uncontrolled hemorrhage induced by 75% tail resection, compared limited resuscitation (target MAP = 40 mmHg) combined with moderate hypothermia (30C) to limited resuscitation at normothermia [65]. Three-day survival rates were significantly greater in hypothermic treated animals (70% vs. 10%). Takasu et al. used the same model to compare the effects of mild (34C) to moderate (30C) hypothermia during limited resuscitation [68]. Once again the greatest survival rate was achieved when limited resuscitation was combined with therapeutic hypothermia (81% vs. 18%). There was no significant difference in survival between the mild and moderate hypothermia groups. In addition, arterial lactate levels were significantly lower in the hypothermia treated animals, suggesting an improved balance of tissue oxygen supply and demand. Prueckner et al. demonstrated similar results but in a pressure controlled model of hemorrhage [69]. While these were elegant studies, the conclusions one can draw are limited as they used a small animal model with minimal physiologic monitoring. Just as important, cooling was accomplished using noninvasive techniques. The latter would be significantly more difficult and would take considerably longer to accomplish in humans or even in a more relevant large animal model. To this end, Wladis et al. used a swine model of severe controlled hemorrhage to examine the effect of therapeutic hypothermia on the hemodynamic response to hemorrhagic shock [70]. The average time to cooling in the hypothermic group was 130 minutes, and temperature in this group ranged from 28.431C. The induction of hypothermia further depressed cardiovascular function and oxygen delivery without increasing mortality. Four-hour survival rates were 73% and 100% in normothermic and hypothermic treated animals respectively. MAP, heart rate, cardiac output, and O2 delivery were significantly lower in hypothermic compared with the normothermic treated animals. In a similar study using the same hemorrhage and hypothermia protocol, Wladis et al. demonstrated that induction of hypothermia during hemorrhagic shock resulted in a significant reduction of adrenaline and noradrenaline concentrations [71]. These authors therefore theorize that the decreases in heart rate, cardiac output and MAP were secondary to a reduction in catecholamine levels, and that this may be cardioprotective. These data are in contrast to other studies of hypothermia, which document a stimulant effect of hypothermia on the sympathetic nervous system [72,73]. The differences between the data from Wladis et al. and that of previous studies may be related to the method of hypothermic induction (i.e., the previous study used immersion techniques) as well as the species of animals studied. These studies suggest that therapeutic hypothermia may provide a mechanism to reduce the imbalance in oxygen

428 Trauma

supply and demand during limited resuscitation, and therefore improve outcome. Although laboratory data appear promising, this therapeutic modality is a long way from becoming a clinical reality. First, we must develop a more complete understanding of the physiologic effects of therapeutic hypothermia, including the mechanism for its protective effects as well as its potential for harmful side effects. Only then will we be able to develop appropriate and safe criteria for its use. In addition, we must develop practical and preferably minimally invasive methods of implementation. In the mean time, therapeutic hypothermia remains somewhat futuristic.

5 6 7

Dillon J, Lynch JL, Myers R, Butcher HR Jr: The treatment of hemorrhagic shock. Surg Gynecol Obstet 1966;122:967977. Wolfman EF Jr, Neill SA, Heaps DK, Zuidemia GD.: Donor blood and isotonic salt solution. Arch Surg 1963;86:869873. Moss GS, Proctor JH, Homer LD, Herman CM, Litt BD.: A comparison of asanguineous fluids and whole blood in the treatment of hemorrhagic shock. Surg Gynecol Obstet 1969;129:12471257. Maughon, JS. An inquiry into the nature of wounds resulting in killed in action in Vietnam. Military Medicine. 1970;135:813. Bellamy RF. The causes of death in conventional land warfare: Implications for combat casualty care research. Military Medicine. 1984;149:5562. Bellamy RF. Death on the battlefield and the role of first aid. Military Medicine. 1987;152:634635. Sauaia A, Moore FA, Moore EE, et al. Epidemiology of trauma deaths: A reassessment. J Trauma 1995;38:185193. Meislin H, Criss E, Judkins D, et al. Fatal Trauma: The modal distribution of time to death is a function of patient demographics and regional resources. J Trauma 1997;43;433440. Trunkey DD, Blaisdell FW. Epidemiology of trauma. Sci Am 4:1, 1988. Gross D, Landau EH, Klin B, et al. Quantitative measurement of bleeding following hypertonic saline therapy in uncontrolled hemorrhagic shock. J Trauma 1989;29(1): 7983. Gross D, Landau EH, Assalia A, et al. Is hypertonic saline resuscitation safe in uncontrolled hemorrhagic shock?. J Trauma 1988;28(6):7516. Sindlinger JF, Soucy DM, Greene SP, et al.. The effects of isotonic saline volume resuscitation in uncontrolled hemorrhage. Surgery, Gynecology & Obstetrics. 1993;177(6): 54550. Smail N, Wang P, Cioffi WG, et al. Resuscitation after uncontrolled venous hemorrhage: Does increased resuscitation volume improve regional perfusion?. J Trauma 1998;44(4):7018. Bickell WH, Bruttig SP, Millnamow GA, et al.. The detrimental effects of intravenous crystalloid after aortotomy in swine. Surgery 1991;110(3): 52936. Kowalenko T, Stern SA, Dronen SC. Improved outcome with hypotensive resuscitation of uncontrolled hemorrhagic shock. J Trauma 1992;33:349 353. Sakles JC, Sena MJ, Knight DA et al. Effect of immediate fluid resuscitation on the rate, volume, and duration of pulmonary vascular hemorrhage in a sheep model of penetrating thoracic trauma. Ann Emerg Med 1997;29(3):392 399. Owens TM, Watson WC, Prough DS, et al. Limiting initial resuscitation of uncontrolled hemorrhage reduces internal bleeding and subsequent volume requirements. J Trauma. 1995;39:200207.

8 9 10 11 12

Conclusions
The most recent studies of hemorrhagic shock and resuscitation suggest that our current protocols are far from ideal, and in fact may be harmful. The data demonstrate that in the setting of uncontrolled hemorrhage, limited or hypotensive resuscitation may be preferable to current guidelines. Contrary to previously held beliefs, substantial periods of moderate hypotension can be tolerated by many trauma patients without progression to end-organ failure. One of the major challenges in resuscitation research will be to develop a mechanism to distinguish those patients who will be able to tolerate such a physiologic stress without major complication. In addition, the data demonstrate that resuscitation methods that yield gradual increases in blood pressure and cardiac output may maximize tissue perfusion without accentuating hemorrhage. And finally, promising new resuscitation modalities, such as HBOCs and therapeutic hypothermia, may provide a method to offset the potential detrimental effects of limited resuscitation. Prior to making sweeping changes in our current treatment guidelines however, randomized controlled clinical trials are necessary. We must not make the mistake of our predecessors and extrapolate data directly from the laboratory to the clinical setting. Hemorrhagic shock from traumatic injury is too complex a disease process, and the trauma population too heterogeneous for us to rely solely on laboratory studies for the answers to our questions.

13 14

15 16

17

18

19

20

21

22

Solomonov E, Hirsh M, Yahiya A, et al. The effect of vigorous fluid resuscitation in uncontrolled hemorrhagic shock after massive splenic injury. Crit Care Med 2000;28(3): 749754. This study compared resuscitation with large volume 0.9% NaCl vs. small volume 7.5% NaCl vs. no resuscitation in a rat model of massive splenic injury, representative of the blunt trauma patient. Aggressive resuscitation significantly increased mortality. The hemodynamic responses to resuscitation in this model mirrored those that have been observed in moderate large vessel injury models. 23 Krausz MM, Bashenko Y, Hirsh M. Crystalloid or colloid resuscitation of uncontrolled hemorrhagic shock after moderate splenic injury. Shock 2000;13: 230235. This study utilized a model of moderate splenic injury in a rat to compare aggressive resuscitation with either crystalloid or colloid to no resuscitation. MAP increased abruptly following initiation of resuscitation, but then rapidly declined to levels equal to or less than that of controls. Hemorrhage volume was significantly greater in the aggressively resuscitated animals. Survival times however, did not differ between groups. 24 Matsuoka T, Hildreth J, Wisner DH. Uncontrolled hemorrhage from parenchymal injury: Is resuscitation helpful? J Trauma 1996;40:915921. Capone AC, Safar P, Stezoski W, et al. Improved outcome with fluid restriction in treatment of uncontrolled hemorrhagic shock. J Am Coll Surg 1995; 180:4956. Stern SA, Dronen SC, Birrer P, et al. The effect of blood pressure on hemorrhage volume and survival in a near-fatal hemorrhage model incorporating a vascular injury. Ann Emerg Med 1993;22:1655163. Stern SA, Kowalenko T, Younger J, et al. Comparison of the effects of bolus vs. slow infusion of 7.5% NaCl/6% Dextran-70 in a model of hear-lethal uncontrolled hemorrhage. Shock 2000;14:616622.

References and recommended reading

Papers of particular interest, published within the annual period of review, have been highlighted as: Of special interest Of outstanding interest 1 Committee on Trauma. Advanced Trauma Life Support Program for Physicians: Instructor Manual. Chicago, Ill: American College of Surgeons; 1993:75110. Wiggers CJ. Experimental hemorrhagic shock. In: Physiology of Shock. New York: Commonwealth Fund; 1950:121146. Shires T, Coln D, Carrico J, Lightfoot S: Fluid therapy in hemorrhage shock. Arch Surg 1964;88:688693. Dillon J, Lynch LJ Jr, Myers R, Butcher HR Jr, Moyer CA. A bioassay of treatment of hemorrhagic shock. I. The roles of blood, Ringers solution with lactate, and macromolecules (dextran and hydroxyethyl starch) in the treatment of hemorrhagic shock in the anesthetized dog. Arch Surgery 1966; 93:53755.

25

26

27

Low-volume fluid resuscitation Stern 429

These investigators compared bolus to slow infusion of HSD in a swine model of near-lethal uncontrolled hemorrhage. Mortality rate and hemorrhage volume were significantly reduced with slow as compared to bolus infusion. The data suggest that outcome from resuscitation of uncontrolled hemorrhage will vary greatly depending on the manner in which a given resuscitation agent is administered. 28 29 Burris D, Rhee P, Kaufmann C, et al. Controlled resuscitation for uncontrolled hemorrhagic shock. J Trauma 1999;46:216223. Krausz MM, Landau EH, Klin B, et al. Hypertonic saline treatment of uncontrolled hemorrhagic shock at different periods from bleeding. Arch Surg 1992;127:9396. Leppaniemi A, Soltero R, Burris D, et al. Fluid resuscitation in a model of uncontrolled hemorrhage: Too much too early, or too little too late? J Surg Research 1996;63;413418. Sixma JJ, Wester J. The hemostatic plug. Semin Hematol 1977;14:265299. Jorgenson L, Borchgrevink CF. The platelet plug in normal persons. Acta Pathol Microbiol Scand 1963;57:4054. Wester J, Sixma JJ, Geuze JJ, et al. Morphology of the hemostatic plug in human skin wounds: Transformation of the plug. Lab Invest 1979;41:182 192. Jorgenson L, Rowsell HC, Hovig T, et al. Resolution and organization of platelet-rich mural thrombi in carotid arteries of swine. Am J Pathol 1967;51:681 693. Hovig T, Stormorken H. Ultrastructural studies on the platelet plug formation in bleeding time wounds form normal individuals and patients with von Willebrands disease. Acta Pathol Microbiol Scand 1974;248(suppl):105122. French JE, Macfarlane RG, Sanders AG. The structure of haemostatic plugs and experimental thrombi in small arteries. Br J Exp Pathol 1964;45:467 474. Bickell WH, Wall MJ, Pepe PE, et al. Immediate versus delayed fluid resuscitation for hypotensive patients with penetrating torso injuries. New Engl J Med 1994;331:11051109. Carrillo P, Takasu A, Safar P, et al. Prolonged severe hemorrhagic shock and resuscitation in rats does not cause subtle brain damage. J Trauma 1998;45: 239249. 58 51 52 53 Bigelow WG, Lindsay W, Greenwood WF: Hypothermia: Its possible role in cardiac surgery. Ann Surg 1950;132:849866. Rosomoff HL, Holaday A: Cerebral blood flow and cerebral oxygen consumption during hypothermia. Am J Physiol 1954;179:8588. Michenfelder JK, Theye RA: The effects of anesthesia and hypothermia on canine cerebral ATP and lactate during anoxia produced by decapitation. Anesthesiology 1970;33:430439. Dempsey RJ, Combs DJ, Maley ME, et al: Moderate hypothermia reduces postischemia edema development and leukotriene production. Neurosurgery 1987;21:177181. Kramer RS, Sanders AP, Lesage AM, et al. The effect of profound hypothermia on preservation of cerebral ATP content during circulatory arrest. J Thorac Cardiovasc Surg. 1968;56:699709. Globus M, Busto R, Lin B, et al. Detection of free radical activity during transient global ischemia and recirculation: effects of intra-ischemic brain temperature modulation. J Neurochem. 1995;65:12501256. Dietrich WD, Alonso O, Busto R, et al. Post-traumatic brain hypothermia reduces histopathologic damage following concussive brain injury in the rat. Acta Neuropathol. 1994;87:250258. Jurkovich GJ, Pitt RM, Curreri PW, Granger DN. Hypothermia prevents increased capillary permeability following ischemia-reperfusion injury. J Surg Res. 1988;44:514521. Luna GK, Maier MV, Palvineg, et al. Incidence and effect of hypothermia in seriously injured patients. J Trauma. 1987;27:10141018. Jurkovich GJ, Greiser WB, Luterman A, et al. Hypothermia in trauma victims: an ominous predictor of survival. J Trauma. 1987;27:10191024. Gentilello LM, Jurkovich GJ, Stark MS, et al. Is hypothermia in the victim of major trauma protective or harmful? Ann Surg 1997;226:439447. Steinemann S, Shackford SR, Davis JW. Implications of admission hypothermia in trauma patients. J Trauma 1990;30:200202. Postel AH, Reid LC, Hinton JW: The therapeutic effect of hypothermia in experimental hemorrhagic shock. Ann Surg 1956;145:311316. Meyer DM, Horton JW: Effect of moderate hypothermia in the treatment of canine hemorrhagic shock. Ann Surg 1988;207:462469. Kim SH, Stezoski W, Safar P, et al. Hypothermia and minimal fluid resuscitation increase survival after uncontrolled hemorrhagic shock in rats. J Trauma 1997;42:213222. Crippen D, Safar P, Porter L, et al. Improved survival of hemorrhagic shock with oxygen and hypothermia in rats. Resuscitation 1991;21:271281. Sori AJ, El-Assuooty A, Rush BF, et al. The effect of temperature on survival in hemorrhagic shock. Am Surg 1987;53:706710. 48 Schultz SC, Powell CC, Burris DG, et al. The efficacy of diaspirin crosslinked hemoglobin solution resuscitation in a model of uncontrolled hemorrhage. J Trauma. 1994;37:408412.

30

Sloan EP, Koenigsberg M, Gens D, et al. Diaspirin cross-linked hemoglobin (DCLHb) in the treatment of severe traumatic hemorrhagic shock: A randomized controlled efficacy trial. JAMA 1999;282:18571864. This was a multicenter, randomized, controlled, single-blinded, efficacy and safety trial of DCLHb in 112 patients with severe traumatic hemorrhagic shock. This was the first trial of a HBOC in patients presenting immediately following injury. This was also the first clinical trial to use the new FDA regulations for waiver of informed consent. The trial was stopped after interim analyses revealed a significantly greater mortality rate in the DCLHb treated patients. 50 Marion DW, Penrod LE, Kelsey SF, et al. Treatment of traumatic brain injury with moderate hypothermia. N Engl J Med 1997;336:540546.

49

31 32 33

34

54

35

55

36

56

37

57

38

Stern SA, Wang X, Mertz M, et al. Under-resuscitation of near-lethal uncontrolled hemorrhage: Effects on mortality and end-organ function at 72 hours. Shock 2001;15:1623. This study evaluated survival and the incidence of end-organ injury 3 days following a 75-minute period of severe (MAP = 40 mmHg) or moderate (MAP = 60 mmHg) under resuscitation of near-lethal uncontrolled hemorrhage in swine. Three-day mortality rates were 36% and 22% in the severely and moderately under resuscitated animals respectively. Surviving animals demonstrated no clinical or histologic evidence of end organ injury at 72 hours. 40 41 Stern SA, Dronen SC, Wang X. A trial of multiple resuscitation regimens in severe hemorrhagic shock. Academic Emergency Medicine 1995;2:8997. Marshall HP, Capone A, Courcoulas AP, et al. Effect of hemodilution on longterm survival in an uncontrolled hemorrhagic shock model in rats. J Trauma 1997;43:673679.

39

59 60 61 62 63 64 65

McNeil JD, Smith LD, Jenkins DH, et al. Hypotensive resuscitation using a polymerized bovine hemoglobin-based oxygen-carrying solution (HBOC201) leads to reversal of anaerobic metabolism. J Trauma 2001;50:1063 1075. In a controlled hemorrhage model, low pressure resuscitation with HBOC-201 is as effective as low pressure resuscitation with whole blood, and more effective than aggressive resuscitation with crystalloid. Manning JE, Katz LM, Brownstein, et al. Bovine hemoglobin-based oxygen carrier (HBOC-201) for resuscitation of uncontrolled, exsanguinating liver injury in swine. SHOCK 2000;13:152159. This study compared low pressure resuscitation with HBOC-201 to that of low pressure resuscitation with RL in a swine liver injury model. The data demonstrate improved survival when limited resuscitation is accomplished with an HBOC as compared to RL. 44 Gulati A, Sen AP. Dose-dependent effect of diaspirin cross-linked hemoglobin on regional blood circulation of severely hemorrhaged rats. Shock. 1998; 9:6573. Schultz SC, Hamilton INJ, Malcolm DS. Use of base deficit to compare resuscitation with lactated Ringers solution, Haemaccel, whole blood, and diaspirin cross-linked hemoglobin following hemorrhage in rats. J Trauma. 1993;35:619625. Cohn SM, Farrell TJ. Diaspirin cross-linked hemoglobin resuscitation of hemorrhage. J Trauma. 1995;39:210216. DeAngeles DA, Scott AM, McGrath AM, et al. Resuscitation from hemorrhagic shock with diaspirin cross-linked hemoglobin, blood, or hetastarch. J Trauma. 1997;42:406412. 43

42

66 67 68

45

Takasu A, Stezoski SW, Stezoski J, et al. Mild or moderate hypothermia, but not increased oxygen breathing, increases long-term survival after uncontrolled hemorrhagic shock in rats. Crit Care Med 2000;28:24652474. This study compared the effects of mild and moderate therapeutic hypothermia to that of normothermia during limited resuscitation of uncontrolled hemorrhage in rats. Survival as well as metabolic parameters were significantly improved with therapeutic hypothermia. There was no difference in outcome between mild and moderate hypothermia treated animals. 69 Prueckner S, Safar P, Rainer K, et al. Mild hypothermia increases survival from severe pressure-controlled hemorrhagic shock in rats. J Trauma 2001;50: 253262. These investigators evaluated the effects of brief and prolonged mild hypothermia as compared to normothermia in a pressure controlled model of hemorrhagic shock. Hypothermia significantly improved survival. Of note is that animals tolerated mild hypothermia for up to 12 hours without complications.

46 47

430 Trauma
Wladis A, Hahn RG, Hjelmqvist H, et al. Acute hemodynamic effects of induced hypothermia in hemorrhagic shock: An experimental study in the pig. SHOCK 2001;15:6064. This study assessed how therapeutic hypothermia modifies the hemodynamic response to hemorrhagic shock in swine. Animals subjected to therapeutic hypothermia experienced greater decreases in blood pressure and cardiac output but a trend toward improved survival as compared to normothermic controls. 71 Wladis A, Hjelmqvist H, Brismar B, et al. Acute metabolic and endocrine effects of induced hypothermia in hemorrhagic shock: An experimental study in the pig. J Trauma 1998;45:527533. 70 72 73 Johnson DG, Hayward JS, Jacobs TP, et al. Plasma norepinephrine responses of man in cold water. J Appl Physiol 1977;43:216220. Hanhela R, Hollmen A, Huttunen P, et al. Plasma catecholamines, corticosterone, glucose, and fatty acid concentrations and mean arterial pressure and body temperature in haemorrhagic hypovolaemia, hypothermia and a combination of these in the rabbit. Acta Physiol Scand 1990;139:441449. Bickell WH, Bruttig SP, Millnomow G, et al. Use of hypertonic saline/dextran versus lactated Ringers solution as a resuscitation fluid after uncontrolled aortic hemorrhage in anesthetized swine. Ann Emerg Med 1992, 21(9):10771085.

74