Severe hemorrhagic shock remains a life threatening event with high mortality. On average, more than 50% of trauma patients with a systolic blood pressure under 90 mm Hg die within the first 24 hours, despite all resuscitative efforts of modern medicine, including blood and catecholamine administration. Among survivors, high rates of infection and the failure of various organ systems impair and determine further prognosis. 1,2
Hemorrhage caused by mechanical trauma or by combat injuries affects the circulatory balance in two ways: it leads to general hypotension and tachycardia, and it causes consecutive peripheral ischemia leading to cellular injury, impairment, and subsequent organ dysfunction, particularly and most severely when appropriate fluid replacement is delayed. 3–6 Prognosis primarily depends upon the grade of hemorrhage, the ischemia tolerance of various organ systems, and, notably, the response time after adequate resuscitation therapy. Overall, the pathophysiologic consequences of hemorrhagic shock involve a very complex cascade of complications that is influenced by additional external factors such as body temperature. Consequently, most therapeutic approaches to hemorrhagic shock are also complex and vary from case to case. 7–11
Mechanical circulatory assist devices (MCADs) have been successfully used in various clinical settings in which cardiac function and peripheral perfusion are impaired. The use of left ventricular assist devices and other MCADs for temporary myocardial and circulatory assistance in the case of acute heart failure (e.g., after acute myocardial infarction or coronary artery bypass surgery) has become a well established management option. 12,13 Left ventricular assist devices and peripheral implantable pumps effectively provide circulatory support in patients with end stage heart failure awaiting heart transplantation (i.e., as bridges to transplantation) and in those requiring long term cardiac assistance (i.e., as bridges to recovery). 14,15 The various assist devices currently available all function by partially or completely augmenting the circulatory performance of the impaired heart or peripheral circulation, restoring normal hemodynamic conditions, and sufficiently perfusing vital end organs.
The extensive knowledge and experience that has been gained from the use of MCADs in patients in acute cardiogenic shock indicates that temporary mechanical assistance can prevent permanent and irreversible organ damage. Additionally, MCAD support has been shown to improve and even normalize end organ function in patients with multisystem failure. 16 Thus we hypothesized that prompt initiation of mechanical circulatory assistance in the event of severe hemorrhagic shock can improve the outcome of this life threatening event.
To test this hypothesis, we conducted a pilot study in which we developed a hemorrhagic shock model in a large animal species (i.e., the pig) leading to delayed death despite pharmacologic intervention within 24 hours and investigated the potential benefit of acute implantation of an MCAD in that model. The primary study endpoint was survival time after hemorrhage. Secondary endpoints were as follows: 1) hemodynamic parameters (e.g., cardiac output [CO], left anterior descending coronary artery [LAD] flow, and superior mesenteric artery [SMA] flow), and 2) biochemical parameters of liver, kidney, and other main organ functions.
Thirteen domestic pigs (mean weight, 55.3 ± 9.7 kg) underwent hemorrhage with or without consecutive pump support. Animal experiments were performed according to the institutional guidelines of the Texas Heart Institute, the Principles of Laboratory Animal Care as formulated by the National Society for Medical Research, and the Guide for the Care and Use of Laboratory Animals issued by the National Institutes of Health (U.S. Department of Health and Human Services, NIH Publication No. 80-23, revised 1985).
After induction and maintenance of anesthesia with a mixture consisting of guaifenesin, ketamine, and xylazine, all pigs underwent surgery including splenectomy. Splenectomy was done because the response to severe hemorrhage of splenectomized swine, as opposed to nonsplenectomized swine, more closely mimics that of humans. 17 Flow probes were attached to the SMA and LAD of each animal. A 12 F double lumen catheter was inserted into the right carotid artery for use later in the hemorrhage inducing procedure (described in the next section). Additionally, complete invasive hemodynamic monitoring was established by inserting arterial and venous pressure lines, including a Swan-Ganz catheter, for monitoring continuous pulmonary artery and wedge pressure. Body temperature was constantly maintained between 35°C and 37°C by using a water blanket connected to a heat exchanger.
A standard centrifugal pump device (BP-80; Medtronic Inc, Minneapolis, MN) was used to provide circulatory assistance in pigs randomized to receive such support (n = 5). Cannulae were inserted through a left thoracotomy into the left atrial appendage and descending thoracic aorta in these pigs.
Hemorrhage and Resuscitation
Hemorrhage was induced by extracting arterial blood from the carotid artery. The pigs were randomized to three experimental hemorrhage groups (Figure 1): group I (n = 3), bleeding to a mean arterial pressure (MAP) of 35 mm Hg; group II (n = 5), bleeding to a MAP of 40 mm Hg; and group III (n = 5), bleeding to a MAP of 40 mm Hg to be followed by MCAD implantation. Once obtained, the target MAP was maintained for 30 minutes. After an additional 30 minutes without any further manipulation, such as hemorrhage or surgery (response time), pigs were then resuscitated by fluid readministration (crystalloid solution/blood loss ratio, 2:1). In group III, MCAD support was initiated 30 minutes after the start of fluid resuscitation. In all cases, blood collected during hemorrhage was saved in heparinized blood bags and used later for autologous blood transfusion during fluid resuscitation.
To maintain comparable hemoglobin levels across groups during the whole study duration, hemoglobin levels were measured hourly. A hemoglobin level lower than 7 g/dl was considered an indication for blood transfusion. If a pig’s supply of autologous blood obtained during hemorrhage was depleted by transfusion, additional blood from a donor pig was administered. Fluid resuscitation was performed by administering physiologic lactate saline (Baxter Healthcare Corp., Deerfield, IL) and hydroxyethyl starch solution (Hespan; B. Braun Medical Inc., Irvine, CA). The ratio of replacement fluids to blood loss during fluid resuscitation was similar in all three study groups: 2 ml of fluids for every 1 ml of blood lost.
To allow quick adjustments in response to metabolic and ventilatory alterations throughout the study, arterial and central venous blood gases (i.e., central venous oxygen saturation; CvO2) were measured hourly. Ventilatory parameters were changed accordingly if necessary. Hemodynamic parameters such as MAP, CO, central venous pressure (CVP), and pulmonary capillary wedge pressure (PCWP) were recorded continuously. For use in measuring end organ function and metabolism, blood samples were collected before hemorrhage (baseline), after hemorrhage, at the beginning of fluid resuscitation, and every 6 hours until death.
All data are presented as mean ± standard deviation (SD). Student’s t-test was used to compare data between groups. Values of p < 0.05 were considered statistically significant.
All pigs survived surgery and subsequent induction of hemorrhage. Pigs in group I were successfully bled to MAP 35 mm Hg; those in groups II and III were bled to MAP 40 mm Hg. A significantly greater volume of bleeding was required to reach the target MAP in group I than in groups II and III (p < 0.05).
After 30 minutes of response time, fluid resuscitation was begun, leading initially to a hyperdynamic circulatory state marked by elevated MAP, LAD flow, SMA flow, and CO levels. The total amounts of fluid administered according to the study protocol within the first 60 minutes of fluid resuscitation were comparable in all three groups: group I, 2,432 ± 652 ml; group II, 2,126 ± 705 ml; and group III, 1,818 ± 586 ml (p = NS).
The mean survival time in group I (230 ± 25.5 min) was significantly shorter than in the other two groups. The longest survival time (22 hours after initiation of resuscitation) was observed in group III. One pig in group III died after only 230 minutes because of ventricular fibrillation after myocardial infarction during pump insertion that led to insufficient filling of the left ventricle and the pump. There was no significant difference in survival time between groups II and III.
Most pigs in all three groups died of ventricular arrhythmia leading to fibrillation refractory to electrical and pharmacologic interventions. In MCAD supported pigs (group III), the arrhythmias led consecutively to insufficient filling of the blood pump and to circulatory failure. In one pig from group III, surgery related failure occurred during implantation of the device, causing ST elevation, myocardial infarction, and finally acute heart failure and premature termination of the experiment (see paragraph above). Another pig in group III received blood from a donor pig 3 hours after hemorrhage, subsequently experienced acute allergic shock, and quickly died.
The total amounts of fluid administered to groups II and III in the posthemorrhagic time period did not significantly differ. Over the course of the study, pigs in group II received a mean total of 11.14 ± 2.2 L of fluid, whereas pigs in group III received a mean total of 9.31 ± 1.6 L (p = NS). Because the survival time of pigs in group I was significantly shorter, the total amounts of fluid administered were markedly smaller as well.
Blood parameters increased progressively in all pigs, indicating the severity of shock. Table 1 gives an overview of selected blood parameters at baseline and after 6 hours for groups II and III. No data are shown for group I because the short survival in that group precluded the capture of long term data and their comparison with data from the other two groups.
In pigs that were bled to MAP 40 mm Hg (groups II and III), statistically significant alterations in creatinine, SGOT, and lactate dehydrogenase levels (among others) were detected between baseline and 6 hours after hemorrhage. Creatinine increased progressively from 1.66 ± 0.4 mg/dl (group II) and 1.44 ± 0.1 mg/dl (group III) before hemorrhage to 2.42 ± 0.5 mg/dl (group II) and 2.2 ± 0.2 mg/dl (group III) 6 hours after initiation of resuscitation efforts (p < 0.05 for each group). SGOT levels in group II increased progressively from 41.2 ± 24.0 U/L at baseline to 750.5 ± 156 U/L 6 hours after hemorrhage (p < 0.05). An attenuated, but still significant, increase in SGOT levels was also observed in group III (from 58 ± 29.7 U/L at baseline to 359.5 ± 181.8 U/L at 6 hours after hemorrhage) (p < 0.05). Changes in SGPT were statistically significant in group II pigs, but not in the MCAD supported pigs of group III. Whereas creatine kinase (CK) and its myocardial isoenzyme CK-MB increased rapidly after induction of hemorrhagic shock in the non-MCAD supported pigs of group II, it remained constant in the MCAD supported pigs of group III.
Changes in CvO2 were similar in groups II and III. Mean SVO2 was 78.6% in group II and 78.8% in group III at baseline but dropped to 46.2% and 47%, respectively, during bleeding. After initiation of fluid resuscitation, SVO2 climbed to 60.8% in group II and 62.2% in group III, and it remained stable thereafter (61% and 60.3%, respectively). In contrast, SVO2 in group I was relatively higher at baseline (89.3 ± 3.9%) but dropped dramatically to 46.8 ± 9.3% at the end of hemorrhage, and it reached levels similar to those in the other two groups at the end of fluid resuscitation (59.1 ± 2.3%).
Overall, hemodynamic data improved after initiation of fluid resuscitation. LAD flow, SMA flow, and CO all remained below baseline levels 1 hour after hemorrhage in group I but returned to or exceeded baseline levels in groups II and III. Thereafter, CO gradually but continuously decreased in groups II and III, although the decrease was smaller in group III (Figure 2). By 8 hours after shock induction, CO in the MCAD supported pigs (group III) fell to 78% of the baseline value (mean reduction, from 5.4 L/min to 4.2 L/min), whereas CO in the non-MCAD supported pigs (group II) fell to 61% of the baseline value (mean reduction, from 6.0 L/min to 3.7 L/min).
LAD and SMA flow after hemorrhage differed significantly among the three groups (Figures 3 and 4). Although hemodynamic data were available for a short time after induction of hemorrhage in group I, there was already a detectable trend in favor of the pigs bled to MAP 40 mm Hg. One hour after initiation of fluid resuscitation, LAD and SMA flow data remained markedly below baseline data before hemorrhagic shock induction. Unfortunately, there were no additional data measurable in this group later on. Although mean LAD flow was lowest in group III at baseline and after hemorrhage, it increased after fluid resuscitation and initiation of MCAD support, eventually achieving statistical significance after 8 hours. Mean LAD flow in group III climbed from 33 ± 5.6 ml/min at baseline to 65 ± 21 ml/min at 8 hours. This increase was significantly greater than the increase observed in group II over the same time interval (39.3 ± 3.6 ml/min) (p < 0.03).
Similar observations were made in terms of SMA flow. Whereas SMA flow in group III had already increased significantly by 1 hour after the end of hemorrhage, it remained relatively constant in group II (88% of baseline at 8 hours).
Right and left ventricular filling pressures, expressed by CVP and PCWP, respectively, were comparable between all three groups at baseline. Before initiation of hemorrhage, invasively measured PCWP was 10.7 ± 7.3 mm Hg in group I, 9.0 ± 4.6 mm Hg in group II, and 9.0 ± 5.1 mm Hg in group III (p = NS). Because of severe intravasal blood loss, PCWP decreased during hemorrhage but returned to baseline (i.e., 10.0 ± 2.5 mm Hg in group I, 9.8 ± 7.5 mm Hg in group II, and 8.6 ± 3.7 mm Hg in group III, respectively) within the first hour of fluid resuscitation. At 5 hours, there was no statistically significant difference between the values for MCAD supported (group III) and non-MCAD supported pigs (group II): 8.0 ± 2.1 mm Hg in group III vs. 10.6 ± 6.7 mm Hg in group II (p = NS).
Of interest is the fact that the baseline data for CVP were similar in all three groups of pigs and returned to baseline within the first hour of fluid resuscitation in pigs bled to MAP 40 mm Hg (3.4 ± 5.0 mm Hg in group II; 5.2 ± 4.0 mm Hg in group III) (p = NS). In contrast, at the same time point, the CVP in group I exceeded the baseline value (10.7 ± 3.1 mm Hg), eventually reaching statistical significance in comparison with group II (p < 0.05) but not in comparison with group III (p = 0.06).
In this pilot study, we developed a porcine hemorrhagic shock model in which the shock was severe enough to cause delayed death within 24 hours, despite fluid resuscitation and various drug interventions, including the use of catecholamines. Pressure fixed bleeding, experimentally elaborated by Wiggers more than 50 years ago, 18 proved to be effective and reliable for our purposes. Bleeding pigs to MAP 35 mm Hg (group I) was found to be too aggressive, leading to expedient death within 2 hours after hemorrhage. Despite fluid resuscitation, CO never recovered in pigs treated in that manner, whereas it returned to baseline in pigs bled to MAP 40 mm Hg (groups II and III).
Although CO increased over time in both groups II and III, there were indications that prompt mechanical circulatory assistance in the event of severe hemorrhagic shock might have an additional beneficial effect. In MCAD supported pigs (group III), coronary and mesenteric blood flow increased significantly from their prehemorrhage levels after fluid resuscitation and initiation of MCAD support; in non-MCAD supported pigs (group II), postresuscitation flow levels dipped progressively after an initial postresuscitation hyperdynamic state. In a sheep model of hemorrhagic shock and fibrillation, Iguidbashian et al.19 noted that 1) sheep supported by cardiopulmonary bypass for 1 hour after shock induction had a significantly better survival rate than sheep resuscitated by epinephrine hydrochloride infusion, and 2) cardiopulmonary bypass preserved left ventricular function, whereas volume resuscitation with or without inotropic support did not.
Treating life threatening, hypovolemic shock with an MCAD that, from a logical point of view, only functions suitably when supplied with a sufficient volume of blood and fluids may seem counterintuitive. However, the rationale for this approach is the device’s potential for overcoming the negative systemic and cardiac effects resulting from the shock induced release of cytokines. Cytokines mediate the postshock development of adult respiratory distress syndrome and multiple organ failure and have negative inotropic effects on the heart. 20–22 These negative inotropic effects are directly treated by MCAD support, which in the process provides sufficient peripheral microcirculation and thereby significantly attenuates the fatal shock cascade. The differences in blood chemistry observed between MCAD supported and non-MCAD supported pigs support this view. Certainly, these aspects need to be investigated further and in more detail.
Various blood parameters changed after bleeding in all three groups. Increases in serum creatinine, liver enzymes, and lactic acid indicated that hemorrhage had critically affected different organ systems. Because survival times for the pigs in group I were so short, there were no long term blood chemistry data available for them. However, this was not a problem in groups II and III. Even though MCAD support in group III could not prevent a general increase in blood parameters, the increase seen in group III was smaller than the one seen in group II (i.e., pigs receiving no MCAD support), suggesting a beneficial effect of MCAD support on end organ functions in this setting. This assumption is supported by the fact that both groups received, on average, the same total amount of fluids. In other words, the difference in blood parameters was more likely caused by the difference in MCAD support than by a dissimilar dilution effect.
Several aspects of the present study were problematic, such as the use of a combination of crystalloid and colloid solutions for fluid resuscitation. To maintain comparable hemoglobin levels across groups, we also administered additional blood from a donor animal when hemoglobin levels dropped below 7 g/dl. Whether crystalloid or colloid solution should be used for fluid resuscitation when blood is not available is a controversial question. Hypercolloidal solutions such as hydroxyethyl starch (15 ml/kg) or large volume Ringer’s lactate may lead to an initial increase in hemodynamic parameters, but, as shown by Krausz et al., 11 such management conversely promotes a significant increase in blood loss with no positive effect on final survival time.
A second problematic aspect, and one that needs to be carefully evaluated in the future, was the implantation site of the cannulae. Manipulation of the heart under shock conditions, when the myocardium is highly irritable, can easily lead to arrhythmia and premature termination of a study. Further, the surgical implantation of the cannulae is highly invasive and poses procedural risks that led to early death in at least one MCAD supported animal. In future experiments, the use of alternative, revolutionary assist devices such as peripheral implantable axial flow pumps, which are quick and easy to implant, highly efficient, and require no thoracotomy access, 23,24 should be considered.
A third problematic aspect was the occurrence of arrhythmias just before death in most of the pigs in this study. This observation may limit the usefulness of the porcine model for studying the potential benefit of volume dependent assist devices in the setting of hemorrhagic shock. Although we analyzed blood gases and electrolytes hourly and immediately corrected any imbalances, ventricular fibrillation was the most common cause of death in this study. In pigs, stressful conditions such as periods of ischemia make the myocardium highly vulnerable and susceptible to ventricular fibrillation and sudden cardiac death. 25–27 In a pig model of myocardial ischemia, McDonald et al.28 found that ventricular fibrillation provokes local release of noradrenaline from sympathetic neurons as part of the sympathetic response. Because increased concentration of catecholamines can ectopically stimulate the heart, it may therefore represent a source of life threatening arrhythmias. One solution to this problem might be to administer antiarrhythmic drugs, whose potential negative inotropic effect could be overcome by the concomitant use of a mechanical assist device.
Naturally, our acute experiments did not allow us to address the long term benefits of circulatory support in terms of preserving main organ (e.g., neurologic) function. However, the fact that alterations of various blood parameters were attenuated under mechanical assistance despite the administration of comparable amounts of fluids in both MCAD and non-MCAD supported pigs is another point in favor of mechanical support. It also suggests that MCAD support in the setting of hemorrhagic shock provides some protection on the cellular level and diminishes the extent of organ dysfunction.
Recently, Shah and colleagues 29 reported that CvO2 seemed to be a reliable predictor of survival in their small animal model of irreversible hemorrhagic shock. In contrast, we found similar SVO2 values in all three groups after hemorrhage despite varying survival times. However, comparisons between different experimental studies on hemorrhagic shock are extremely difficult and depend upon the comparability of the treatment protocols in terms of parameters such as mode and site of hemorrhage or resuscitation management. Therefore, we developed our own large animal model, which as a bonus enables studies with different MCADs.
Body temperature is an important factor in the efficacy of resuscitative efforts in the event of hemorrhagic shock. Recently, in a porcine hemorrhagic shock model, Krause et al.9 found that aggressive volume replacement restored CO to baseline levels in normothermic pigs but depressed it in hypothermic pigs. They concluded that shock and hypothermia have an additive, deleterious effect in combination and that both conditions need to be aggressively addressed when they occur simultaneously. In the present study, we successfully treated the hypothermia induced by our surgical procedures (which included both laparatomy and thoracotomy) by using a heating pad, thereby achieving normothermic conditions throughout the experiment.
In conclusion, we have developed a good, reproducible, large animal model of prolonged hemorrhagic shock and shown that the application of mechanical circulatory support in this model improves perfusion and leads to less profound alterations in end organ (i.e., kidney and liver) function. Though the long term effects (especially the long term preservation of end organ functions) require further investigation, our model represents a novel and potentially beneficial approach to preventing or curtailing the harmful consequences of prolonged hemorrhagic shock.
1. Heckbert SR, Vedder NB, Hoffman W, et al: Outcome after hemorrhagic shock in trauma patients. J Trauma 45: 545–549, 1998.
2. Craig RL, Poole GV: Resuscitation in uncontrolled hemorrhage. Am Surg 60: 59–62, 1994.
3. Schlag G, Redl H, Hallstrom S: The cell in shock: the origin of multiple organ failure. Resuscitation 21: 137–180, 1991.
4. Shoemaker WC, Peitzman AB, Bellamy R, et al: Resuscitation from severe hemorrhage. Crit Care Med 24 (Suppl): S12–S23, 1996.
5. Bellamy RF, Maningas PA, Wenger BA: Current shock models and clinical correlations. Ann Emerg Med 15: 1392–1395, 1986.
6. Cutress R: Fluid resuscitation in traumatic haemorrhage. J Accid Emerg Med 12: 165–172, 1995.
7. Falk JL, O’Brien JF, Kerr R: Fluid resuscitation in traumatic hemorrhagic shock. Crit Care Clin 8: 323–340, 1992.
8. Haizlip TM, Poole GV, Falzon AL: Initial resuscitation volume in uncontrolled hemorrhage: effects on organ function. Am Surg 65: 215–217, 1999.
9. Krause KR, Howells GA, Buhs CL, et al: Hypothermia-induced coagulopathy during hemorrhagic shock. Am Surg 66: 348–354, 2000.
10. Mizushima Y, Wang P, Cioffi WG, Bland KI, Chaudry ICH: Should normothermia be restored and maintained during resuscitation after trauma and hemorrhage? J Trauma 48: 58–65, 2000.
11. Krausz MM, Bashenko Y, Hirsh M: Crystalloid or colloid resuscitation of uncontrolled hemorrhagic shock after moderate splenic injury. Shock 13: 230–235, 2000.
12. Golding LA: Postcardiotomy mechanical support. Semin Thorac Cardiovasc Surg 3: 29–32, 1991.
13. DeRose JJ Jr, Umana JP, Argenziano M, et al: Improved results for postcardiotomy cardiogenic shock with the use of implantable left ventricular assist devices. Ann Thorac Surg 64: 1757–1762, 1997.
14. Frazier OH, Rose EA, McCarthy P, et al: Improved mortality and rehabilitation of transplant candidates treated with a long-term implantable left ventricular assist system. Ann Surg 222: 327–336, 1995.
15. Goldstein DJ, Oz MC, Rose EA: Medical progress: implantable left ventricular assist devices. N Engl J Med 339: 1522–1533, 1998.
16. Burnett CM, Vega JD, Radovancevic B, et al: Improved survival after Hemopump insertion in patients experiencing postcardiotomy cardiogenic shock during cardiopulmonary bypass. ASAIO Trans 36: M626–M629, 1990.
17. Hannon JP, Bossone CA: The conscious pig as a large animal model for studies of hemorrhagic hypotension, in Tumbleson ME (ed), Swine in Biomedical Research
, vol. 3, New York: Plenum Press, 1986, p. 1424.
18. Wiggers CJ: The present status of the shock problem. Physiol Rev 22: 74–123, 1942.
19. Iguidbashian JP, Follette DM, Contino JP, Chao C, Berkoff HA: Improved myocardial function using cardiopulmonary support in resuscitation for hemorrhagic shock. Arch Surg 129: 1013–1017, 1994.
20. Roumen RM, Hendriks T, van der Ven-Jongekrijg J, Nieuwenhuijzen GAP: Cytokine patterns in patients after major vascular surgery, hemorrhagic shock, and severe blunt trauma relation with subsequent adult respiratory syndrome and multiple organ failure. Ann Surg 218: 769–776, 1993.
21. Bahrami S, Yao YM, Leichtfried G, Redl H, Marzi I, Schlag G: Significance of TNF in hemorrhage-related hemodynamic alterations, organ injury, and mortality in rats. Am J Physiol 272: H2219–H2226, 1997.
22. Finkel MS, Oddis CV, Jacob TD, Watkins SC, Hattler BG, Simmons RL: Negative inotropic effects of cytokines on the heart mediated by nitric oxide. Science 257: 387–389, 1992.
23. Reitan O, Sternby J, Ohlin H: Hydrodynamic properties of a new percutaneous intra-aortic axial flow pump. ASAIO J 46: 323–329, 2000.
24. Siess T, Nix C, Menzler F: From a lab type to a product: a retrospective view on Impella’s assist technology. Artif Organs 25: 414–421, 2001.
25. Parker GW, Michael LH, Entman ML: An animal model to examine the response to environmental stress as a factor in sudden cardiac death. Am J Cardiol 60: 9J–14J, 1987.
26. Bharati S, Levine M, Huang SK, et al: The conduction system of the swine heart. Chest 100: 207–212, 1991.
27. Ovize M, Aupetit JF, Rioufol G, et al: Preconditioning reduces infarct size but accelerates time to ventricular fibrillation in ischemic pig heart. Am J Physiol 269: H72–H79, 1995.
28. McDonald FM, Knopf H, Hartono S, et al: Acute myocardial ischaemia in the anaesthetised pig: local catecholamine release and its relation to ventricular fibrillation. Basic Res Cardiol 81: 636–645, 1986.
29. Shah NS, Kelly E, Billiar TR, et al: Utility of clinical parameters of tissue oxygenation in a quantitative model of irreversible hemorrhagic shock. Shock 10: 343–346, 1998.