When severe blood loss accompanies trauma, delaying intravenous resuscitation until bleeding is controlled can reduce morbidity and mortality compared to the immediate application of fluid therapy (1). A major rationale for such under resuscitation is the observation that aggressive resuscitation raises arterial pressure thereby accentuating the loss of red cells and plasma proteins from the sites of bleeding (2–5). However, a delay in the restoration of blood volume may place poorly perfused organs at risk for cellular injury by prolonging ischemia and by reperfusion when full resuscitation is initiated. Several pharmacologic measures to attenuate the physiologic mechanisms that contribute to such injury are under active investigation (6–12). However, the influence of the rate of resuscitation after hemostasis has been established on end-organ damage and ultimate outcome remains to be determined.
A previous study by Lilly and colleagues (13) examined the influence of resuscitation rate after nonlethal hemorrhage in the dog. Although rapid intravenous infusions of crystalloid restored blood volume more quickly than a slow continuous infusion, the rapid infusions also suppressed the hormonal responses that act to reestablish homeostasis (14) in the face of hypovolemia. Furthermore, the restoration of blood volume and cardiac output after 90 min of the slow infusion was complete when the volume infused in the slow group was only half that in a group that received its infusion as a 10 min bolus (13). At 3 h when the intravenous resuscitation was complete in all groups, the recovery of cardiac stroke volume and heart rate was better after the slow infusion than after the bolus administration of fluid (13). The present study was designed to examine this issue in unanesthetized rats that had experienced hypovolemia of a magnitude and duration such that they were at risk both for death and for hepatocellular damage.
MATERIALS AND METHODS
All procedures and protocols were approved by the Institutional Animal Care and Use Committee of the University of Maryland School of Medicine. The experiments were performed in adherence to the National Institutes of Health Guidelines on the Use of Laboratory Animals. Male Sprague-Dawley rats (283-416 g) were anesthetized with ketamine hydrochloride (Phoenix Pharmaceutical, St. Joseph, MO; 75 mg/kg, i.p.) and xylazine hydrochloride (Phoenix; 5 mg/kg, i.p.). Femoral arterial and venous catheters were placed using aseptic procedures (15,16). All incisions were filled with Bacitracin Ointment (E. Fougera & Co., Melville, NY) and 2% Xylocaine Jelly (Astra Pharmaceutical Products, Inc., Westborough, MA) to prevent infection and desensitize the wound, respectively. Incisions were then closed using 3-0 silk interrupted sutures. All catheters were tunneled subcutaneously to exit at the back of each animal's neck. The catheters were then passed through a stainless steel spring to exit at the top of each cage. Rats were housed singly in a temperature and humidity controlled environment with lights on at 0700 h and off at 1900 h each day. Catheters were flushed daily. Care was taken to minimize activity in the animal quarters during these procedures. Experiments were conducted on the fourth post-operative day.
Experimental rationale and protocol
In preliminary experiments we attempted to bleed the awake rats to a fixed mean arterial pressure (MAP) of 40 mmHg for 90 min until resuscitation as has been done in acutely prepared anesthetized rats by others (9). However, the defense of arterial pressure on the fourth post-operative day in the conscious rat was so potent that more than 39 mL/kg had to be withdrawn to maintain the hypotension. Rats that entered a reuptake phase died quickly when fluid was given to reverse the hypotension, and the remaining rats died when resuscitation was initiated at 90 min. In subsequent studies we found that a two-stage hemorrhage of 36 mL/kg total volume put chronically prepared rats at risk for hepatocellular damage and death whereas hemorrhage of 33 mL/kg produced little such risk. However, the rats with the lowest arterial pressures after hemorrhage of 36 mL/kg were also at risk for early death that did not respond to fluid administration. We arrived at the following protocol that had good survival for 2.5 h and a significant risk of later death. On the day of the experiment, each rat was randomly assigned to receive one of four different resuscitation protocols. After baseline hemodynamic parameters and blood samples were obtained, water bottles were removed, and the unanesthetized rats were bled 33–36 mL/kg from the arterial catheter. Hemorrhage was done in two stages. Initially, 24 mL/kg (40% blood volume) was withdrawn in 4–8 min. Thirty minutes after the onset of the first hemorrhage, 1 mL less than 12 mL/kg was drawn in 3 to 5 min. During the next 5 min an additional 0.5 mL was withdrawn if the rat's MAP stabilized to a value >60 mmHg. This procedure was repeated if MAP stabilized again to a value >60 mmHg. Thus, rats that were able to defend their MAP above 60 mmHg during the first 45 min of the experiment were bled a total of 36 mL/kg. Rats with less potent compensation of arterial pressure received a total hemorrhage that was 0.5 or 1 mL less than this volume. The possible total volumes of hemorrhage in a rat weighing 333 g were 36, 34.5, and 33 mL/kg. No further blood was withdrawn after 45 min.
Hypovolemia was maintained for an additional 105 min for a total hypovolemic period of 150 min from the onset of the first hemorrhage. If MAP fell below 45 mmHg during the last 100 min of the hypovolemic period, Lactated Ringer's solution (LR, Baxter Healthcare Corp., Deerfield, IL) was administered 0.5 to 1 mL at a time to prevent further hypotension. This rescue volume was subtracted from the total infusate volume used for resuscitation. At the end of the total hypovolemic period (150 min), four modes of intravenous fluid resuscitation were studied: LR (3× hemorrhage volume minus any rescue volume) over 5 min (FAST), 20 min (MEDIUM), 180 min (SLOW), and no infusion of LR (NO RESUS). Food and water were allowed ad libitum at the end of each infusion of LR and 2.5 h after hemorrhage in the NO RESUS group. Animals were followed for 72 h with collection of hemodynamic data and blood samples at assigned time points.
Assays and statistical analysis
Blood samples of 250 μL were collected using (ethylenedianitrilo)tetraacetic acid as an anticoagulant. Plasma was separated by centrifugation and the samples were stored at −70°C until analyzed. Plasma ornithine carbamoyltransferase (OCT) was measured using the method of Ohshita (17) modified to use one-tenth of the stated volume so that samples of 20 μL could be assayed. The enzymatic activity of the plasma was determined from the ability of each sample to catalyze the production of citrulline from substrate. One unit of activity is defined as 1 μmole of citrulline per min at 37°C. Plasma creatinine and lactate were determined colorimetrically with commercially available kits (Creatinine Kit #555 and Lactate Kit #735, Sigma Diagnostics Inc., St. Louis, MO). However, the acidification step that corrects for substances that crossreact in the creatinine assay was omitted because the addition of acid to rat plasma made it turbid. Sensitivities for the various assays were as follows: the OCT assay 1 IU/L, the creatinine assay 0.7 mg/dl, and the lactate assay 4 mg/dl.
Differences in the proportions of survivors were tested as described by Natrella (18). Hemodynamic data and plasma OCT responses were tested with analysis of variance (ANOVA) corrected for repeated measures in time (19). Because the distribution of the OCT data was non-Gaussian, the statistical analysis was done on the square roots of the raw values. This monotonic transformation improved both the distribution and homogeneity of variance of the OCT data to comply with the assumptions that underlie ANOVA. The means reported are derived solely from survivors. Unequal cell frequencies were analyzed by the method of unweighted means based on the random assignment of the rats to the various treatments and no a priori expectation of mortality rates (19). When significant F values were obtained simple main effects within treatment groups or at particular sample times were determined with the Newman-Keuls procedure (19). P < 0.05 was considered statistically significant. All data have been expressed as mean ± standard error of the mean.
Baselines and fluid volumes
Of 61 rats that were hemorrhaged, 54 (88.5%) survived until full resuscitation was initiated at 2.5 h after the onset of the first hemorrhage. Table 1 shows mean data for the rats treated with each resuscitation protocol. There were no significant differences between treatments in mean body weight, total hemorrhage volume, and the amount of rescue fluid administered to prevent hypotension <50 mmHg during the hypovolemic period. The resuscitation volume administered at 2.5 h did not differ between the resuscitated groups.
MAP and heart rate (HR) both before and during the hypovolemic period did not differ among the four tx groups (Figs. 1 and 2). MAP decreased after each episode of hemorrhage (P < 0.01, Fig 1) and showed significant recovery after the first hemorrhage but not after the second hemorrhage. HR only decreased after the initial hemorrhage (P < 0.01, Fig. 2) and then increased to values similar or greater than the baseline before the first hemorrhage. At 10 min into resuscitation, MAP increased in the MEDIUM tx group to 82 ± 4.4 mmHg to exceed the values in all other groups (P < 0.05, Fig. 1). MAP in the MEDIUM group then declined by 30 min after the resuscitation to a value that was not different from those during the hypovolemic period after the second hemorrhage. Although MAP initially increased after the FAST infusion, the response was transient and was not maintained even for the duration of the infusion. The FAST infusion did elicit a transient bradycardia at 10 min from the onset of resuscitation (P < 0.05, Fig. 2). At 1 h after resuscitation and 3.5 h after hemorrhage MAP in the FAST group declined to a value that was less than the values in the SLOW and MEDIUM tx groups but did not differ from the value in the NO RESUS group. From 5 to 72 h after the initial hemorrhage, MAP and HR in all surviving rats did not differ among groups. MAP increased slowly during this time with significant increases from the minimal values during the hypovolemic period at 24 h after SLOW and FAST tx, at 48 in the NO RESUS group, and at 72 h after MEDIUM tx.
Plasma OCT was undetectable in a majority of rats before and at the end of the hypovolemic period. Significant increases were not seen until full resuscitation was initiated. The response occurred earliest after FAST treatment when OCT increased to a value greater than the values in all other groups by 30 min after the resuscitation was initiated (at 3 h in Fig. 3). OCT in the FAST group remained elevated compared to the other treatments for the first 24 h. In the groups that did not receive the FAST infusion significant increases in OCT from baseline were not observed until 24 h. All increases in OCT resolved in rats that survived for 3 days from the hypovolemic period.
Plasma lactate increased after hemorrhage to levels that did not differ between groups before the onset of resuscitation (P < 0.01, Fig. 4). At 30 min after the onset of resuscitation plasma lactate in the FAST group attained a value greater than the values in the remaining groups (P < 0.05). Lactate recovered to values that did not differ from baseline at 5.5 h in the NO RESUS and SLOW groups and at 7.5 h in the MEDIUM and FAST groups.
Plasma creatinine did not change significantly in any group before resuscitation (Fig. 5). However, creatinine at 7.5 and 24 h was less in all groups that received intravenous fluid than in the NO RESUS group. This trend continued but was not significant at 48 h.
Figure 6 shows the survival in all groups. Most deaths occurred during the first 3 h of resuscitation in all but the MEDIUM tx group for which deaths occurred after 30 h. Survival at 72 h was less in the FAST (57%) and No tx groups (58%) than in the SLOW (87%) and MEDIUM tx groups (85%), p < 0.05.
Until recently, the cornerstone of management in trauma-induced hemorrhage has been the early restoration of circulating blood volume. However, recent experimental (2,4,20–22) and clinical (1) data indicate that it may be best to delay or limit the administration of intravenous fluids in hypovolemic patients until hemostasis is achieved. Delaying full resuscitation could reduce the blood flow to certain organs due to hypotension or compensatory vasoconstriction. These organs might then be vulnerable to cellular damage as a consequence of either reduced flow or reperfusion when full resuscitation is instituted. Our results show that the rate of resuscitation after an extended period of hypovolemia influences hepatocellular integrity and survival.
All measured variables were within normal limits at baseline and did not differ between groups during the hypovolemic period (Table 1 and Figs. 1, 2, 4). Furthermore, the total volume of hemorrhage and the rescue volume required to prevent excessive hypotension before full resuscitation was initiated were similar among all groups (Table 1). Thus, we have no evidence that the observed differences in survival and the hepatic release of OCT resulted from factors other than those associated with the different rates of resuscitation.
Although end-organ injury is a well-known consequence of hemorrhage in anesthetized animals to a constant level of hypotension (Wigger's prep) followed by resuscitation, our findings were obtained in chronically prepared, unanesthetized, animals. Anesthesia, by definition, provides hypnosis with systemic analgesia and blunts central autonomic reflexes and responses to stimuli. The volume of hemorrhage in the present study was greater than that used in acutely prepared, anesthetized rats to maintain MAP at ∼40 mmHg (7,9,10). However, the rats in the present study defended their arterial pressure above the level used in the acute studies but were still at risk for hepatocellular injury and mortality.
The hypotension and bradycardia that followed the first episode of hemorrhage showed some recovery from 15 to 30 min as has been observed previously in this preparation (23). The second episode of hemorrhage then reversed the recovery of arterial pressure with no further decrease in the heart rate. Greater hypotension in the FAST group than in the other groups from 1.5 to 2.5 h after hemorrhage (Fig. 1) may have influenced the later effects of FAST resuscitation. However, MAP at all time points before resuscitation and the volumes of rescue and hemorrhage in the FAST group were not significantly different from those in the remaining groups.
Although the rate of resuscitation influenced survival in the present study, the effect of this rate on the hemodynamic responses was modest and of short duration. At 10 min into resuscitation, MAP increased transiently in the MEDIUM group. Transient hypertension and overshoot in cardiac output that parallel an overexpansion of blood volume have been reported in dogs that are infused with normal saline at a similar rate after hemorrhage (13). Thus, the transient increase in MAP after the MEDIUM infusion was likely the result of such an overshoot in volume during the infusion of LR. In contrast, no increase in MAP was seen at 10 min after the FAST infusion. Although MAP in both the MEDIUM and the FAST groups began to increase at the onset of the infusions, peak responses that were not maintained occurred in both groups before all of the prescribed volume had been administered (detailed data not shown). At 60 min into resuscitation, the SLOW and MEDIUM treatments led to better maintenance of MAP than did the FAST infusion for which MAP was significantly less. Note that this response in the SLOW group occurred when it had received only 33% of the volume given to the other resuscitated groups. Despite these early differences sustained recovery of MAP was not observed until at least 24 h in all groups. Given the improved survival in the SLOW and MEDIUM groups it appears that delayed, hypotensive resuscitation shown previously to be beneficial in uncontrolled hemorrhage is also efficacious in controlled hemorrhage.
Hemorrhage and resuscitation are known to elicit a decrease in cardiac contractility (24) that may have contributed to the failure of resuscitation to restore MAP during the early period. This response may have been aggravated by fluid overload particularly after the FAST infusions that elicited transient bradycardia (Fig. 2) at 10 min after the onset of resuscitation. Increased cardiac filling time and rapid volume expansion could enlarge the heart to a region of the Starling relationship with a flat or negative slope and account for the earliest deaths in the FAST group. Fluid overload as well as diffusion of LR from the vascular space might also contribute to the transient nature of the MAP response to the MEDIUM rate of infusion. However, the possible lethal effect of such overload appears to be overridden by a benefit that prevents early cardiovascular collapse in the face of prolonged hypovolemia. Failure to negate the progressive effects of hypovolemia is the likely cause of the early death after NO RESUS or SLOW tx.
The release of the enzyme OCT indicated that hepatocellular damage occurred after all treatments. However, OCT increased earliest after the FAST infusions with a significant increase at 30 min after the start of resuscitation. The rapid infusion of the entire resuscitation volume in this group may have accelerated the washout of this hepatic enzyme. However, the infusion of LR was also complete in the MEDIUM group at 20 min after the start of resuscitation, and no significant increase in OCT occurred after the MEDIUM infusion until 24 h. Thus, it appears most likely that the FAST infusion accelerated the onset of hepatocellular damage relative to the other treatments. Although tissue ischemia is a prerequisite for such damage, the reoxygenation that occurs with reperfusion is an important initiating event. The specific mechanisms that underlie the accelerated damage in the FAST treatment group have yet to be determined but some possibilities deserve attention. Firstly, rapid resuscitation may lead to rapid reperfusion of ischemic tissue with an abrupt increase in tissue oxygen tension. Such an event, even if transient, may accentuate the production of reactive oxygen species. In particular, the irreversible production of peroxynitrite through the reaction of radicals with nitric oxide induced by hemorrhagic shock has been implicated as a cause of cellular injury (25). Reactive oxygen species also stimulate the accumulation of neutrophils (11,26), and hemorrhage and resuscitation may elicit mitochondrial defects that trigger cell death (27). Secondly, LR has been reported to have adverse inflammatory actions (28) compared to other volume expanders. It is possible that these effects are accentuated by the FAST rate of infusion.
OCT in the remaining groups increased at 24 h. Survivors at this time in the NO RESUS group were observed to drink with the return of their water bottles and had likely replaced some of their volume deficit. Whether the greater release of OCT at 24 h in the NO RESUS group compared to the SLOW group resulted from an extended period of reduced hepatic blood flow in NO RESUS group or a difference in reperfusion from the SLOW group is not known. Nonetheless, our results suggest that the MEDIUM and SLOW infusions were optimal to reintroduce oxygen and deliver LR at rates that minimize tissue damage.
Another risk of delayed resuscitation is excessive and prolonged accumulation of lactate (29). During the hypovolemic period plasma lactate increased in all of the treatment groups. The high mean value of lactate before resuscitation in the FAST group may have contributed to its greater mortality compared to the other resuscitated groups. However, the difference between all of the groups at this time was not statistically significant (P > 0.1). This result contrasted with that at 30 min after the onset of resuscitation when lactate was greater after the FAST infusion than after the other treatments (P < 0.05). At 3 h after the onset of resuscitation, lactate had decreased significantly in all groups even though the exogenous administration of LR in the resuscitated groups opposed this effect. The return of plasma lactate to its baseline before hemorrhage occurred at 3 h after resuscitation in the SLOW and NO RESUS groups as opposed to 5 h in the MEDIUM and FAST groups. Thus, the SLOW infusion of LR proved to be the optimal intravenous therapy for reversing the elevation of this metabolite.
Unlike lactate, plasma creatinine did not increase during the hypovolemic period. However, the infusion of fluid elicited decreases in this metabolite such that the value after 5 h of intravenous resuscitation at any rate was less than that after no resuscitation (Fig. 5). Thus, fluid resuscitation appears to improve renal function with no advantage as the rate of infusion is increased. Because the values of creatinine in all of the groups were only modestly elevated, our data indicate that renal failure is not an important component in this experimental model. This result may reflect the degree of hypotension before resuscitation. Greater increases in creatinine occur in shock models with hypovolemia that reduces MAP to ≤40 mmHg and where the mortality exceeds 80% (8,12). Our findings are consistent with the observation of Stern and colleagues (22) that under resuscitation during an uncontrolled hemorrhage to maintain MAP at 60 mmHg with a mortality of 22% minimizes renal dysfunction but does not eliminate the increase in plasma transaminases.
Survival at 72 h after the FAST infusion was no better than it was in the group that was not resuscitated. Furthermore, the survival after these two treatments was reduced compared to survival after the SLOW and MEDIUM infusions. Although the survival at 72 h was nearly identical in the SLOW and MEDIUM groups, there was a difference in the time at which death occurred. In the SLOW group, death occurred within 3 h of the onset of the infusion before all of the prescribed fluid was administered. Accordingly, it appeared that SLOW tx failed to correct the hemorrhage-induced hypovolemia quickly enough to prevent cardiovascular collapse. In contrast, survival after MEDIUM tx was 100% for more than one day after the resuscitation was complete. Thus, the MEDIUM infusions appeared to prevent the immediate adverse effects of the hypovolemia but did not eliminate late lethal complications. The incidence of mortality in these two groups was not sufficient to determine the significance of these observations.
These results suggest an optimal pattern for the administration of intravenous fluid. In the initial phase of resuscitation, sufficient volume could be infused at a moderate rate to prevent cardiovascular collapse. Then the rate of infusion could be reduced to support further restoration of volume through both exogenous and endogenous mechanisms. Future investigation is warranted to determine the indicators and the parameters for such a strategy that would optimize survival and minimize cellular damage in various scenarios of hemorrhagic shock.
1. Bickell WH, Wall MJ, Pepe PE, Martin RR, Ginger VF, Allen MK, Mattox KL: Immediate versus delayed fluid resuscitation for hypotensive patients with penetrating torso injuries. N Engl J Med 331:1105–1109, 1994.
2. Burris D, Rhee P, Kaufmann C, Pikoulis E, Austin B, Eror A, DeBraux S, Guzzi L, Leppaniemi A: Controlled resuscitation for uncontrolled hemorrhagic shock. J Trauma 46:216–223, 1999.
3. Capone A, Safar P, Stezoski SW, Peitzman A, Tisherman S: Uncontrolled hemorrhagic shock outcome model in rats. Resuscitation 29:143–152, 1995.
4. Owens TM, Watson WC, Prough DS, Uchida T, Kramer GC: Limiting initial resuscitation of uncontrolled hemorrhage reduces internal bleeding and subsequent volume requirements. J Trauma 39:200–209, 1995.
5. Stern SA, Dronen SC, Birrer P, Wang X: Effect of blood pressure on hemorrhage volume and survival in a near-fatal hemorrhage model incorporating a vascular injury. Ann Emerg Med 22:155–163, 1993.
6. Harbrecht BG, Wu B, Watkins SC, Marshall HP, Pietzman AB, Billiar TR: Inhibition of nitric oxide synthase during hemorrhagic shock increases hepatic injury. Shock 4:332–337, 1995.
7. Harbrecht BG, Wu B, Watkins SC, Billiar TR, Pietzman AB: Inhibition of nitric oxide synthesis during severe shock but not following resuscitation increases hepatic injury and neutrophil accumulation in hemorrhaged rats. Shock 8:415–421, 1997.
8. Hua TC, Moochhala SM: Influence of L-arginine, aminoguanidine, and NGnitro-L-arginine methyl ester (L-name) on the survival rate in a rat model of hemorrhagic shock. Shock 11:51–57, 1999.
9. Menezes J, Hierholzer C, Watkins SC, Lyons V, Peitzman AB, Billiar TR, Tweardy DJ, Harbrecht BG: A novel nitric oxide scavenger decreases liver
injury and improves survival after hemorrhagic shock. Am J Physiol 277:G144–G151, 1999.
10. Wattanasirichaigoon S, Menconi MJ, Fink MP: Lisofylline ameliorates intestinal and hepatic injury induced by hemorrhage and resuscitation in rats. Crit Care Med 28:1540–1549, 2000.
11. Yamakawa Y, Takano M, Patel M, Tien N, Takada T, Bulkley GB: Interaction of platelet activating factor, reactive oxygen species generated by xanthine oxidase and leukocytes in the generation of hepatic injury after shock/resuscitation. Ann Surg 231:387–398, 2000.
12. Zhong Z, Enomoto N, Connor HD, Moss N, Mason RP, Thurman RG: Glycine improves survival after hemorrhagic shock in the rat. Shock 12:54–62, 1999.
13. Lilly MP, Gala GJ, Carlson DE, Sutherland BE, Gann DS: Saline resuscitation after fixed-volume hemorrhage: Role of resuscitation volume and rate of infusion. Ann Surg 216:161–171, 1992.
14. Gann DS, Ward DG, Carlson DE: Neural control of ACTH: a homeostatic reflex. Recent Prog Horm Res 34:357–400, 1978.
15. Carlson DE, Babus JK, Nguyuza N, Melhem-Stancofski H, Eastridge BJ: Role of endotoxin in the response to experimentally induced bacteremia in chronically prepared rats. Am J Physiol 272:R1562–R1570, 1997.
16. Darlington DN, Keil LC, Dallman MF: Potentiation of hormonal responses to hemorrhage and fasting, but not hypoglycemia in conscious adrenalectomized rats. Endocrinology 125:1398–1406, 1989.
17. Ohshita M, Takeda H, Kamiyama Y, Ozawa K, Honjo I: A direct method for the estimation of ornithine carbamoyltransferase activity in serum. Clinica Chimica Acta 67:145–152, 1976.
18. Natrella MG:Experimental Statistics.
Washington, D.C.: U. S. Govt. Printing Office, 1966.
19. Winer BJ:Statistical Principles in Experimental Design.
New York: McGraw-Hill, 1971.
20. Leppaniemi A, Soltero R, Burris D, Pikoulis E, Waasdorp C, Ratigan J, Hufnagel H, Malcolm D: Fluid resuscitation in a model of uncontrolled hemorrhage: too much too early, or too little too late? J Surg Res 63:413–418, 1996.
21. Riddez L, Johnson L, Hahn RG: Central and regional hemodynamics during crystalloid fluid therapy after uncontrolled intra-abdominal bleeding. J Trauma 44:433–439, 1998.
22. Stern SA, Wang X, Mertz M, Chowanski ZP, Remick DG, Kim HM, Dronen SC: Under-resuscitation of near-lethal uncontrolled hemorrhage: effects on mortality and end-organ function at 72 hours. Shock 15:16–23, 2001.
23. Darlington DN, Chew G, Ha T, Keil LC, Dallman MF: Corticosterone, but not glucose, treatment enables fasted adrenalectomized rats to survive moderate hemorrhage. Endocrinology 127:766–772, 1990.
24. Horton JW: Hemorrhagic shock depresses myocardial contractile function in the guinea pig. Circ Shock 28:23–35, 1989.
25. Szabo C: The pathophysiological role of peroxynitrite in shock, inflammation, ischemia-reperfusion injury. Shock 6:79–88, 1996.
26. Morita Y, Clemens MG, Miller LS, Rangan U, Kondo S, Miyasaka M, Yoshikawa T, Bulkley GB: Reactive oxidants mediate TNF-α-induced leukocyte adhesion to rat mesenteric venular endothelium. Am J Physiol 269:H1833–H1842, 1995.
27. Lemasters JJ: Mechanisms of hepatic toxicity V. Necrapoptosis and the mitochondrial permeability transition: shared pathways to necrosis and apoptosis. Am J Physiol 276:G1–G6, 1999.
28. Rhee P, Burris D, Kaufmann C, Pikoulis M, Austin B, Ling G, Harviel D, Waxman K: Lactated ringer's solution resuscitation causes neutrophil activation after hemorrhagic shock. J Trauma 44:313–319, 1998.
29. Baron BJ, Sinert RH, Sinha AK, Buckley MC, Shaftan GW, Scalea TM: Effects of traditional versus delayed resuscitation on serum lactate
and base deficit. Resuscitation 43:39–46, 1999.