Sensitivity to endotoxin in rabbits is increased after hemorrhagic shock WILLIAM
J. MILESKI,
ROBERT
K. WINN,
JOHN
Departments of Surgery, Medicine, and Physiology-Biophysics, School of Medicine, Seattle, Washington 98104 MILESKI, WILLIAM AND CHARLES L. RICE.
J., ROBERT
K.
WINN,
JOHN
M.
HARLAN,
Sensitivity to endotoxin in rabbits is increased after hemorrhagic shock. J. Appl. Physiol. 73(3): 1146-
1149, 1992.-The immunoinflammatory response following trauma and hemorrhage may predisposeto the development of sepsisand multiple-organ failure syndrome. Cardiac output (CO), arterial pressure,arterial PO,, and pulmonary permeability index were measured.We examined the sensitivity of rabbits to infusions of lipopolysaccharide (LPS) after hemorrhagic shock. Shock wasproduced by reducing CO to 40% of baseline for 90 min, followed by resuscitation with shedblood and then with lactated Ringer solution to maintain CO near baseline. Animals were assignedto three groups: I) hemorrhagic shock only, 2) LPS only, and 3) hemorrhagic shock + LPS. Groups 1 and 3 were subjected to hemorrhagic shock on day 1. Escherichia coZi LPS wasinfused (1.0 pg/kg iv) into groups 2 and 3 on day 2. Fluid resuscitation with lactated Ringer solution was continued in an effort to maintain CO at baseline. Five hours after LPS infusion, 1251-albuminwas injected intravenously, and rabbits were killed 1 h later for measurementof pulmonary permeability index. LPS infusion after shock (group 3) caused significant decreasesin CO, arterial pressure, and PO, and an increase in pulmonary permeability. These changeswere not seen in the groups 1 and 2. We conclude that hemorrhagic shock and resuscitation result in a proinflammatory state, leading to increasedsensitivity to subsequentexposure to LPS.
M. HARLAN,
AND
University
CHARLES
L. RICE
of Washington
duces many of the physiological effects of sepsis in animals and in humans (13, 18, 19). The deleterious effects of the interaction of hemorrhagic shock and LPS have not been clearly defined. In this study we examined the effects of moderate hemorrhagic shock combined with small doses of LPS to determine whether hemorrhagic shock causes an increased sensitivity to the effects of LPS. We chose to separate hemorrhage and LPS administration by 24 h under the assumption that in a clinical setting patients would not be exposed to significant LPS for some period after their traumatic episode. METHODS
New Zealand White rabbits (1.5-2.0 kg) were divided into three experimental groups, and experiments were performed over a 2-day period. Animals were anesthetized with ketamine (20-40 mg/kg iv) supplemented with local lidocaine. Then, with sterile techniques, catheters were placed in the vena cava and aorta via the internal jugular vein and carotid artery, respectively. The aortic catheter had a thermistor at its tip for measurement of temperature and determination of cardiac output (CO) by thermal dilution. These catheters remained in place sepsis; multiple-organ failure; hemorrhagic shock; lung for 24 h and were carefully wrapped with a tube bandage to protect the catheter from being dislodged and to prepermeability vent the rabbits from chewing on them. Baseline measurements of CO, temperature, mean arterial pressure (MAP), white blood cell (WBC) counts, hematocrit, and MULTIPLE-ORGAN FAILURE SYNDROME (MOFS) is recog- arterial blood gases were obtained after recovery from anesthesia. Group 1 animals were subjected to hemornized as a leading cause of late mortality in trauma patients and may result from an overstimulated immurhagic shock on day 1, followed by 6 h of observation, and noinflammatory system. However, the complex immunoon day 2 they were again monitored for 6 h without furinflammatory response that predisposes to the developther manipulation. Group 2 rabbits were monitored for 6 ment of organ failure is not completely understood (5,7). h on day 1; they were given Escherichia coli LPS (1 pg/kg) Recent investigations have suggested that synergistic on day 2 and followed for 6 h. Group 3 rabbits were substimuli act to produce a progressive dysfunction and, ultijected to hemorrhagic shock on day 1 and followed for 6 mately, failure of multiple organs (11). Severe trauma h; on day 2 they were given E. coli LPS (1 pg/kg) and and sepsis are important risk factors for the development followed for 6 h. of MOFS, and the combination of the two may act synerHypovolemic shock was induced and maintained for 90 gistically to produce MOFS. We have postulated that hy- min by withdrawing blood from the venous catheter into potension may be causal in organ injury following sterile heparinized syringes. CO was reduced to 40% of trauma and have shown experimentally that hemorbaseline and maintained at that level by further blood rhagic shock produces organ failure (12, 20, 21). Also, withdrawal. During the shock period, CO, MAP, and tembacterial endotoxin or lipopolysaccharide (LPS), the perature were monitored at 15min intervals. The withdrawn blood was anticoagulated with heparin (10 U/ml) principal toxin of the cell wall of Gram-negative bacteria, and remained at room temperature during the shock pehas been implicated as a major mediator of sepsis. It pro1146
0161-7567192
$2.00
Copyright 0 1992 the American Physiological
Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.
SEN
-
0
2
Sl’l’l
V I’I’Y
‘I’U
EN LWl’WXlN
Shock LPS Shock+LPS
4
Time
0246
(hours)
FIG. 1. Cardiac output in animals subjected to shock only (shock, n = 3), animals subjected to sham hemorrhage followed by lipopolysaccharide (LPS, n = 6), and animals subjected to hemorrhage and LPS (shock + LPS, n = 6). Values are means -t SE. Shock + LPS group was significantly different from the other 2 groups at 5 and 6 h by analysis of variance and Scheffe’s F test (P < 0.05).
riod. Shed blood was reinfused over -30 min at the conclusion of the 90-min shock period, and monitoring continued for 6 h. CO, MAP, and temperature were measured at hourly intervals. Fluid resuscitation with lactated Ringer solution was continued throughout the 6-h period to maintain CO and MAP near baseline levels. At the conclusion of the resuscitation period, the arterial and venous catheters were flushed with saline containing heparin (100 U/ml) and secured with the tube bandage, and animals were returned to their cages. Rabbits had free access to food and water overnight. On the following morning ( -24 h postshock), monitoring was resumed and fluid resuscitation conti .nued as described above. Groups 2 and 3 were given LPS on experimental day 2. The LPS was diluted in 10 ml of sterile saline at a concentration to result in a dose of 1.0 pg/kg and infused intravenously over 30 min, Fluid resuscitation with lactated Ringer solution was continued for the next 6 h. Animals were killed 6 h after LPS injection (groups 2 and 3) or after 6 h of observation (group 1) on day 2. One hour before death, all animals were given an intravenous injection of -5 &i of 1251-labeledalbumin, a dose similar to that used by others (22). Postmortem bronchoalveolar lavage was performed using 20 ml of saline to lavage the left lung. The lung was lavaged in and out four times with this volume to better sample the alveolar space. The lavage volume was chosen such that a large percentage of the initial volume was recovered. The 1251-albuminin the lavage effluent and in plasma was measured in a gamma counter, and the 1251-albuminlavage-to-plasma ratio was calculated as an index of pulmonary permeability (7). Statistical analysis was performed by analysis of variance and Scheffe’s test, with significance at P < 0.05. Data in the text are means t SD. Values in Figs. l-4 are means t SE for convenience of display.
AF’l’l!X
Hl3MUfCfCHAtilC
decrease in CO, systemic pressure, or arterial oxygenation. The animal in group 3 died as a result of cardiovascular collapse, with a last measured CO -30% of baseline and MAP of 44 Torr. CO for all three groups over the &day experimental period are shown in Fig. 1. There were no differences in the baseline CO among groups. The severity and duration of the hemorrhagic shock measured by decreased CO between the animals in groups 1 and 3 were not different. On day 2, group 3 had decreased CO during the last 2 h after LPS injection. These decreases were statistically significant (P < 0.05) compared with groups 1 and 2. Reduction in CO occurred, even with significant fluid resuscitation. Groups 1 and 2 maintained CO at near-baseline levels throughout the monitoring period and were not statistically different from each other. Averaged arterial PO, for all three experimental groups are presented in Fig. 2 for the baseline value on each experimental day and for the final value for each animal. There were no differences among the groups during day 1 of observation. On day 2, however, group 3 developed profound hypoxemia, and the averaie final measurement was statistically different by analysis of variance and Scheffe’s test (P < 0.05). Hemorrhagic shock alone or LPS alone did not result in altered oxygenation, inasmuch as arterial PO, remained near baskline. Averaged WBC counts for all three groups are presented in Fig. 3. Baseline WBC counts were similar for the three groups. Hemorrhagic shock on day 1 (groups 1 and 3) produced a significant leukocytosis after resuscitation from shock that returned to preshock levels by the start of day 2. WBC counts in group 1 did not change from baseline on day 2. Group 2 showed a transient decrease in circulating WBC counts after infusion of LPS. In contrast, group 3 developed a prolonged and severe leukopenia that reached a nadir 3 h after LPS injection and persisted throughout the observation period. The averaged WBC counts were statistically significantly different 1, 2, and 5 h after LPS infusion. The fluid requirements to maintain CO and MAP near baseline on day 2 were statistically significantly different q Shock Cl Shock+LPS 100.0 c 5 z
c 75.0 0 .m m c $ 50.0 s 0) * 25.0 K
0
0.0 RESULTS
Two animals died as a result of LPS infusion: one in group 2 and one in group 3. There was no apparent physiological change in the animal in group 2: there was no
SHWClK
Day 1 BL
FIG. 2. Arterial PO, in (n = 6) groups. Values are was significantly different variance and Scheffe’s F
Day 2 BL
Final
shock (n = 3), LPS (n = 6), and shock + LPS means & SE. Final PO, in shock + LPS group from that in the other 2 groups by analysis of test (P < 0.05). BL, baseline.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.
1148
SENSITIVITY
TO
ENDOTOXIN
among the three groups by analysis of variance (P < 0.05). Groups 1 and 2 needed little fluid on day 2: 6.4 t 3.1 and 10.1 t 4.2 ml/kg, respectively. Group 3 was given 36.5 t 12.3 ml/k g, h owever, CO still decreased, even with this substantial fluid resuscitation. Group 3 was statistically different from groups 1 and 2 (P < 0.05), but groups 1 and 2 were not different from each other by Scheffe’s test. The averaged 12’1-albumin lavage-to-plasma ratio for the three groups and the average of four normal animals killed 1 h after injection of 1251-albumin are presented in Fig. 4. There was a >25fold increase in the 1251-albumin lavage-to-plasma ratio for group 3 compared with groups 1 and 2 and the normal animals. There were no significant differences among normal animals, group 1, and group 2; however, these three groups were different from group 3 by analysis of variance and Scheffe’s test l
(P < 0.05). DISCUSSION
The pathogenesis of MOFS remains elusive, despite intensive investigation into the cascade of events following trauma and hemorrhagic shock. A number of investigators have proposed that the development of MOFS after shock and trauma is related to host immunosuppression and increased susceptibility to infection and sepsis (1,6,8,15). Although this may be true, organ failure in rabbits and nonhuman primates as a result of hemorrhagic shock can be prevented by blocking neutrophil function with a monoclonal antibody directed to the CD18 leukocyte adhesion complex (12,21). Protection by this anti-inflammatory monoclonal antibody is consistent with its blocking the effects of an overstimulated inflammatory system. This leads to an alternate hypothesis that MOFS can occur as a result of a “primed” inflammatory system followed by an otherwise insignificant septic challenge. In this study we have demonstrated that hemorrhage caused an increased susceptibility of the cardiovascular and pulmonary systems to the adverse effects of endotoxin. This was manifested by decreased CO, decreased arterial oxygenation, and increased pulmonarv permeability. These adverse effects
+ -
Shock Shock LPS
00
2
4
+ LPS
I - I - 1 - 1 0 2 4 6
Time (hours) FIG. 3. Leukocyte (WBC) number in shock (n = 3), LPS (n = 6), and shock + LPS (n = 6) groups. Values are means & SE. Groups were statistically different from each other at 1, 2, and 5 h on Clay 2 by analysis of variance (P < 0.05).
AFTER
HEMORRHAGIC
SHOCK
0.100
3 iG z
0.075
L Q) n
.
0.000
- h-----4
Control
- b-----7n
Shock
= brr-r-r’
LPS
///,// k\\\\\\ ,,//,, .\.\\\\ ,,,,// \\\\\NX ,//,/, k\\\\\\ ,,,/// .\\\\\\ - -----*
Shock
’
+ LPS
FIG. 4. Permeability index (1251-albumin alveolar lavage-to-plasma ratio) for shock (n = 3), LPS (n = 4), shock + LPS (n = 3), and normal animals (control, n = 4). Animals were killed 1 h after injection of 1251-albumin and alveolar lavage was performed. Values are means t SE. Shock + LPS group was statistically different from all other groups, but the other groups were not different from each other by analysis of variance and Scheffe’s F test (P < 0.05).
cannot be attributed to an inability to recognize or kill bacteria, because LPS was used as the second stimulus. Similar findings have been reported after burn injury. Burns are associated with an increased clinical incidence of MOFS, and experimental burns have been shown to increase the sensitivity to LPS (14). This increased sensitivity to LPS might be important in the pathogenesis of MOFS. The precise mechanism(s) that lead to the deleterious effects of LPS are not completely defined. LPS does not cause a direct toxic effect, at least at low concentrations, but the host immunoinflammatory system might be overstimulated, resulting in injury to the host as a result of LPS (10). Mononuclear phagocytes and the vascular endothelium may be responsible for the host response to LPS. In vitro experiments have shown that LPS causes activation of mononuclear phagocytes, leading to production of cytokines such as tumor necrosis factor (TNF) and interleukin-1. These cytokines cause activation of endothelial cells, leading to expression of a proinflammatory and procoagulant phenotype. In addition, the in vivo deleterious effects of LPS are mimicked at least in part by TNF, and the effects of LPS can be reduced by administration of anti-TNF antibody (2). Thus it is possible that part of the interaction of hemorrhagic shock and LPS administration may result from stimulation or “priming” of monocytes as a result of hemorrhagic shock. The pathophysiological changes that occur after hemorrhagic shock leading to the increased sensitivity to intravenously administered LPS are undetermined. Several possibilities present themselves. Hemorrhagic shock may lead to stimulation and/or priming of the host macrophage population, resulting in increased production of cytokines, both constitutively and in response to LPS. This possibility is supported by the recent demonstration of enhanced Kuppfer cell production of interleukin-1 and TNF in the first 24 h after shock (1). Neutrophils have been implicated as important media-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.
SENSITIVITY
TO
ENDOTOXIN
tors of organ injury in both shock and sepsis, and their adherence is thought to be critical in causing an injury. The profound and prolonged leukopenia after LPS administration in animals subjected to shock and LPS may be the result of proadhesive alterations resulting from cytokines released during or shortly after shock. The endothelium is known to become proadhesive as a result of surface expression of adhesion molecules after stimulation with cytokines (3, 9). Increased endothelial adherence molecule expression in response to sepsis but not hemorrhagic shock has been reported (16). Thus, if the number of adhesion molecules on the endothelial surface is increased by LPS after hemorrhage compared with no hemorrhage, it would be the result of priming by hemorrhage rather than direct stimulation. Hemorrhagic shock has also been associated with gastrointestinal tract injury and subsequent breakdown of the gastrointestinal mucosal barrier, allowing translocation of bacteria and bacterial products (4,17). Thus hemorrhage may cause a transient bacteremia or endotoxemia, leading to a generalized Shwartzman reaction. The generalized Shwartzman reaction is produced by intravenous injection of LPS followed by a second intravenous dose 18-24 h later. It has remained an immunopathological curiosity for over 60 years, and it is possible that hemorrhage followed by sepsis is the clinical analogue of a generalized Shwartzmann reaction. We conclude that hemorrhagic shock and resuscitation result in a proinflammatory state by 24 h after shock rather than a suppressed state, and this leads to increased sensitivity to the subsequent exposure to LPS. Understanding this increased sensitivity to LPS may contribute to a better understanding of the pathogenesis of MOFS and allow for development of therapeutic drugs for the treatment of this life-threatening syndrome. This study was supported by National Institutes of Health Grants HL-43141, HL-30542, GM-07037, and GM-42686 and by the American Heart Association of Washington. This study was approved by the Animal Care Committee of the University of Washington, and all experiments were performed in accordance with National Institutes of Health guidelines for the use of laboratory animals. Address for reprint requests: R. K. Winn, Dept. of Surgery ZA-16, Harborview Medical Center, 325 9th Ave., Seattle, WA 98104. Received
17 December
1990; accepted
in final
form
21 April
1992.
REFERENCES 1. AYALA, A., M. M. PERRIN, AND I. H. CHAUDRY. Increased susceptibility to sepsis following hemorrhage: defective Kupffer cell-mediated antigen presentation. Surg. Forum 40: 102-104, 1989. 2. BEUTLER, B., I. W. MISLARK, AND A. C. CERAMI. Passive immunization against cachectin/tumor necrosis factor protects mice from lethal effect of endotoxin. Science Wash. DC 234: 370-373, 1985. 3. BEVILACQUA, M., J. POBER, M. WHEELER, R. COTRAN, AND M.
AFTER
HEMORRHAGIC
SHOCK
1149
GIMBRONE. Interleukin 1 acts on cultured human vascular endothelium to increase the adhesion of polymorphonuclear leukocytes, monocytes, and related leukocyte cell lines. J. CLin. Inuest. 76: 2003-2010, 1985. 4. CAMERINI, D., S. JAMES, I. STAMENKOVICH, AND B. SEED. Leu-81 TQl is a human equivalent of MEL-14 lymph node homing receptor. Nature Lond. 342: 78-82, 1989. 5. CARRICO, C. J., J. L. MEAKINS, J. C. MARSHALL, D. FRY, AND R. V. MAIER. Multiple organ failure syndrome. Arch. Surg. 121: 196-208, 1986. 6. ESRIG, B. C., L. FRAZEE, S. F. STEPHENSON, H. C. POLK, R. L. FULTON, AND C. E. JONES. The predisposition to infection following hemorrhagic shock. Surg. Gynecol. Obstet. 144: 915-917, 1977. 7. FAIST, E., A. E. BAUE, H. DITTMER, AND G. HEBERER. Multiple organ failure in polytrauma patients. J. Trauma 23: 775-787, 1983. 8. FAIST, E., A. MEWES, T. STRASSER, S. WALZ, C. BAKER, W. ERTEL, AND G. HBERER. Alteration of monocyte function following major injury. Arch. Surg. 123: 287-292, 1988. 9. GAMBLE, J., J. HARLAN, S. KLEBANOFF, AND M. VADAS. Stimulation of the adherence of neutrophils to umbilical vein endothelium by human recombinant tumor necrosis factor. Proc. Natl. Acad. Sci. USA 82: 8667-8671, 1985. 10. LARRICK, J. W. Antibody inhibition of the immunoinflammatory cascade. J. Crit. Care 4: 211-224, 1989. 11. MEAKINS, J. L. Etiology of multiple organ failure. J. Trauma 30: S165S168, 1990. 12. MILESKI, W. J., R. K. WINN, N. V. VEDDER, T. H. POHLMAN, J. M. HARLAN, AND C. L. RICE. Inhibition of CD18-dependent neutrophil adherence reduces organ injury after hemorrhagic shock in primates. Surgery 108: 205-212, 1990. 13. MORRISON, D. C. Bacterial endotoxins and pathogenesis. Rev. Infect. Dis. 5: S733-S747, 1983. 14. NERLICH, M., J. FLYNN, AND R. H. DEMLING. Effect of thermal injury on endotoxin-induced lung injury. Surgery 93: 289-296, 1983. 15. POLK, H. C. The enhancement of host defenses against infection. Search for the holy grail? Surgery 99: l-6, 1986. 16. REDL, H., G. SCHLAG, H. P. DINGES, W. BURRMAN, R. ROTHLEIN, J. POBER, AND R. COTRAN. Endothelial activation after polytrauma and sepsis in the baboon (Abstract). Circ. Shock 31: 79, 1990. 17. SORI, A. J., B. F. RUSH, T. W. LYSZ, S. SMITH, AND G. W. MACHIEDO. The gut as source of sepsis after hemorrhagic shock. Am. J. Surg. 155: 187-192, 1988. 18. SIJFFREDINI, A. F., R. E. FROMM, M. M. PARKER, M. BRENNER, J. A. KOVACS, R. A. WESLEY, AND J. E. PARRILLO. The cardiovascular response of normal humans to the administration of endotoxin. N. Engl. J. Med. 321: 280-287, 1989. 19. VANDEVTNER, S. J. H., H. R. BULLER, J. W. CATE, A. STURK, AND W. PAUW. Endotoxemia: an early predictor of septicemia in febrile patients. Lancet 1: 605-609, 1988. 20. VEDDER, N. B., B. W. FOUTY, R. K. WINN, J. M. HARLAN, AND C. L. RICE. Role of neutrophils in generalized reperfusion injury associated with resuscitation from shock. Surgery 106: 509-516, 1989. 21. VEDDER, N. B., R. K. WINN, C. L. RICE, E. CHI, K.-E. ARFORS, AND J. M. HARLAN. A monoclonal antibody to the adherence promoting leukocyte glycoprotein CD18 reduces organ injury and improves survival from hemorrhagic shock and resuscitation in rabbits. J. Clin. Inuest. 81: 939-944, 1988. 22. WORTHEN, G. S., C. HASLETT, A. J. REES, R. S. GUMBAY, J. E. HENSON, AND P. M. HENSON. Synergistic effect of trace amounts of lipopolysaccharide and neutrophil stimuli on vascular permeability and neutrophil sequestration in the lung. Am. Rev. Respir. Dis. 136: 19-28, 1987.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.
J. MILESKI,
ROBERT
K. WINN,
JOHN
Departments of Surgery, Medicine, and Physiology-Biophysics, School of Medicine, Seattle, Washington 98104 MILESKI, WILLIAM AND CHARLES L. RICE.
J., ROBERT
K.
WINN,
JOHN
M.
HARLAN,
Sensitivity to endotoxin in rabbits is increased after hemorrhagic shock. J. Appl. Physiol. 73(3): 1146-
1149, 1992.-The immunoinflammatory response following trauma and hemorrhage may predisposeto the development of sepsisand multiple-organ failure syndrome. Cardiac output (CO), arterial pressure,arterial PO,, and pulmonary permeability index were measured.We examined the sensitivity of rabbits to infusions of lipopolysaccharide (LPS) after hemorrhagic shock. Shock wasproduced by reducing CO to 40% of baseline for 90 min, followed by resuscitation with shedblood and then with lactated Ringer solution to maintain CO near baseline. Animals were assignedto three groups: I) hemorrhagic shock only, 2) LPS only, and 3) hemorrhagic shock + LPS. Groups 1 and 3 were subjected to hemorrhagic shock on day 1. Escherichia coZi LPS wasinfused (1.0 pg/kg iv) into groups 2 and 3 on day 2. Fluid resuscitation with lactated Ringer solution was continued in an effort to maintain CO at baseline. Five hours after LPS infusion, 1251-albuminwas injected intravenously, and rabbits were killed 1 h later for measurementof pulmonary permeability index. LPS infusion after shock (group 3) caused significant decreasesin CO, arterial pressure, and PO, and an increase in pulmonary permeability. These changeswere not seen in the groups 1 and 2. We conclude that hemorrhagic shock and resuscitation result in a proinflammatory state, leading to increasedsensitivity to subsequentexposure to LPS.
M. HARLAN,
AND
University
CHARLES
L. RICE
of Washington
duces many of the physiological effects of sepsis in animals and in humans (13, 18, 19). The deleterious effects of the interaction of hemorrhagic shock and LPS have not been clearly defined. In this study we examined the effects of moderate hemorrhagic shock combined with small doses of LPS to determine whether hemorrhagic shock causes an increased sensitivity to the effects of LPS. We chose to separate hemorrhage and LPS administration by 24 h under the assumption that in a clinical setting patients would not be exposed to significant LPS for some period after their traumatic episode. METHODS
New Zealand White rabbits (1.5-2.0 kg) were divided into three experimental groups, and experiments were performed over a 2-day period. Animals were anesthetized with ketamine (20-40 mg/kg iv) supplemented with local lidocaine. Then, with sterile techniques, catheters were placed in the vena cava and aorta via the internal jugular vein and carotid artery, respectively. The aortic catheter had a thermistor at its tip for measurement of temperature and determination of cardiac output (CO) by thermal dilution. These catheters remained in place sepsis; multiple-organ failure; hemorrhagic shock; lung for 24 h and were carefully wrapped with a tube bandage to protect the catheter from being dislodged and to prepermeability vent the rabbits from chewing on them. Baseline measurements of CO, temperature, mean arterial pressure (MAP), white blood cell (WBC) counts, hematocrit, and MULTIPLE-ORGAN FAILURE SYNDROME (MOFS) is recog- arterial blood gases were obtained after recovery from anesthesia. Group 1 animals were subjected to hemornized as a leading cause of late mortality in trauma patients and may result from an overstimulated immurhagic shock on day 1, followed by 6 h of observation, and noinflammatory system. However, the complex immunoon day 2 they were again monitored for 6 h without furinflammatory response that predisposes to the developther manipulation. Group 2 rabbits were monitored for 6 ment of organ failure is not completely understood (5,7). h on day 1; they were given Escherichia coli LPS (1 pg/kg) Recent investigations have suggested that synergistic on day 2 and followed for 6 h. Group 3 rabbits were substimuli act to produce a progressive dysfunction and, ultijected to hemorrhagic shock on day 1 and followed for 6 mately, failure of multiple organs (11). Severe trauma h; on day 2 they were given E. coli LPS (1 pg/kg) and and sepsis are important risk factors for the development followed for 6 h. of MOFS, and the combination of the two may act synerHypovolemic shock was induced and maintained for 90 gistically to produce MOFS. We have postulated that hy- min by withdrawing blood from the venous catheter into potension may be causal in organ injury following sterile heparinized syringes. CO was reduced to 40% of trauma and have shown experimentally that hemorbaseline and maintained at that level by further blood rhagic shock produces organ failure (12, 20, 21). Also, withdrawal. During the shock period, CO, MAP, and tembacterial endotoxin or lipopolysaccharide (LPS), the perature were monitored at 15min intervals. The withdrawn blood was anticoagulated with heparin (10 U/ml) principal toxin of the cell wall of Gram-negative bacteria, and remained at room temperature during the shock pehas been implicated as a major mediator of sepsis. It pro1146
0161-7567192
$2.00
Copyright 0 1992 the American Physiological
Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.
SEN
-
0
2
Sl’l’l
V I’I’Y
‘I’U
EN LWl’WXlN
Shock LPS Shock+LPS
4
Time
0246
(hours)
FIG. 1. Cardiac output in animals subjected to shock only (shock, n = 3), animals subjected to sham hemorrhage followed by lipopolysaccharide (LPS, n = 6), and animals subjected to hemorrhage and LPS (shock + LPS, n = 6). Values are means -t SE. Shock + LPS group was significantly different from the other 2 groups at 5 and 6 h by analysis of variance and Scheffe’s F test (P < 0.05).
riod. Shed blood was reinfused over -30 min at the conclusion of the 90-min shock period, and monitoring continued for 6 h. CO, MAP, and temperature were measured at hourly intervals. Fluid resuscitation with lactated Ringer solution was continued throughout the 6-h period to maintain CO and MAP near baseline levels. At the conclusion of the resuscitation period, the arterial and venous catheters were flushed with saline containing heparin (100 U/ml) and secured with the tube bandage, and animals were returned to their cages. Rabbits had free access to food and water overnight. On the following morning ( -24 h postshock), monitoring was resumed and fluid resuscitation conti .nued as described above. Groups 2 and 3 were given LPS on experimental day 2. The LPS was diluted in 10 ml of sterile saline at a concentration to result in a dose of 1.0 pg/kg and infused intravenously over 30 min, Fluid resuscitation with lactated Ringer solution was continued for the next 6 h. Animals were killed 6 h after LPS injection (groups 2 and 3) or after 6 h of observation (group 1) on day 2. One hour before death, all animals were given an intravenous injection of -5 &i of 1251-labeledalbumin, a dose similar to that used by others (22). Postmortem bronchoalveolar lavage was performed using 20 ml of saline to lavage the left lung. The lung was lavaged in and out four times with this volume to better sample the alveolar space. The lavage volume was chosen such that a large percentage of the initial volume was recovered. The 1251-albuminin the lavage effluent and in plasma was measured in a gamma counter, and the 1251-albuminlavage-to-plasma ratio was calculated as an index of pulmonary permeability (7). Statistical analysis was performed by analysis of variance and Scheffe’s test, with significance at P < 0.05. Data in the text are means t SD. Values in Figs. l-4 are means t SE for convenience of display.
AF’l’l!X
Hl3MUfCfCHAtilC
decrease in CO, systemic pressure, or arterial oxygenation. The animal in group 3 died as a result of cardiovascular collapse, with a last measured CO -30% of baseline and MAP of 44 Torr. CO for all three groups over the &day experimental period are shown in Fig. 1. There were no differences in the baseline CO among groups. The severity and duration of the hemorrhagic shock measured by decreased CO between the animals in groups 1 and 3 were not different. On day 2, group 3 had decreased CO during the last 2 h after LPS injection. These decreases were statistically significant (P < 0.05) compared with groups 1 and 2. Reduction in CO occurred, even with significant fluid resuscitation. Groups 1 and 2 maintained CO at near-baseline levels throughout the monitoring period and were not statistically different from each other. Averaged arterial PO, for all three experimental groups are presented in Fig. 2 for the baseline value on each experimental day and for the final value for each animal. There were no differences among the groups during day 1 of observation. On day 2, however, group 3 developed profound hypoxemia, and the averaie final measurement was statistically different by analysis of variance and Scheffe’s test (P < 0.05). Hemorrhagic shock alone or LPS alone did not result in altered oxygenation, inasmuch as arterial PO, remained near baskline. Averaged WBC counts for all three groups are presented in Fig. 3. Baseline WBC counts were similar for the three groups. Hemorrhagic shock on day 1 (groups 1 and 3) produced a significant leukocytosis after resuscitation from shock that returned to preshock levels by the start of day 2. WBC counts in group 1 did not change from baseline on day 2. Group 2 showed a transient decrease in circulating WBC counts after infusion of LPS. In contrast, group 3 developed a prolonged and severe leukopenia that reached a nadir 3 h after LPS injection and persisted throughout the observation period. The averaged WBC counts were statistically significantly different 1, 2, and 5 h after LPS infusion. The fluid requirements to maintain CO and MAP near baseline on day 2 were statistically significantly different q Shock Cl Shock+LPS 100.0 c 5 z
c 75.0 0 .m m c $ 50.0 s 0) * 25.0 K
0
0.0 RESULTS
Two animals died as a result of LPS infusion: one in group 2 and one in group 3. There was no apparent physiological change in the animal in group 2: there was no
SHWClK
Day 1 BL
FIG. 2. Arterial PO, in (n = 6) groups. Values are was significantly different variance and Scheffe’s F
Day 2 BL
Final
shock (n = 3), LPS (n = 6), and shock + LPS means & SE. Final PO, in shock + LPS group from that in the other 2 groups by analysis of test (P < 0.05). BL, baseline.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.
1148
SENSITIVITY
TO
ENDOTOXIN
among the three groups by analysis of variance (P < 0.05). Groups 1 and 2 needed little fluid on day 2: 6.4 t 3.1 and 10.1 t 4.2 ml/kg, respectively. Group 3 was given 36.5 t 12.3 ml/k g, h owever, CO still decreased, even with this substantial fluid resuscitation. Group 3 was statistically different from groups 1 and 2 (P < 0.05), but groups 1 and 2 were not different from each other by Scheffe’s test. The averaged 12’1-albumin lavage-to-plasma ratio for the three groups and the average of four normal animals killed 1 h after injection of 1251-albumin are presented in Fig. 4. There was a >25fold increase in the 1251-albumin lavage-to-plasma ratio for group 3 compared with groups 1 and 2 and the normal animals. There were no significant differences among normal animals, group 1, and group 2; however, these three groups were different from group 3 by analysis of variance and Scheffe’s test l
(P < 0.05). DISCUSSION
The pathogenesis of MOFS remains elusive, despite intensive investigation into the cascade of events following trauma and hemorrhagic shock. A number of investigators have proposed that the development of MOFS after shock and trauma is related to host immunosuppression and increased susceptibility to infection and sepsis (1,6,8,15). Although this may be true, organ failure in rabbits and nonhuman primates as a result of hemorrhagic shock can be prevented by blocking neutrophil function with a monoclonal antibody directed to the CD18 leukocyte adhesion complex (12,21). Protection by this anti-inflammatory monoclonal antibody is consistent with its blocking the effects of an overstimulated inflammatory system. This leads to an alternate hypothesis that MOFS can occur as a result of a “primed” inflammatory system followed by an otherwise insignificant septic challenge. In this study we have demonstrated that hemorrhage caused an increased susceptibility of the cardiovascular and pulmonary systems to the adverse effects of endotoxin. This was manifested by decreased CO, decreased arterial oxygenation, and increased pulmonarv permeability. These adverse effects
+ -
Shock Shock LPS
00
2
4
+ LPS
I - I - 1 - 1 0 2 4 6
Time (hours) FIG. 3. Leukocyte (WBC) number in shock (n = 3), LPS (n = 6), and shock + LPS (n = 6) groups. Values are means & SE. Groups were statistically different from each other at 1, 2, and 5 h on Clay 2 by analysis of variance (P < 0.05).
AFTER
HEMORRHAGIC
SHOCK
0.100
3 iG z
0.075
L Q) n
.
0.000
- h-----4
Control
- b-----7n
Shock
= brr-r-r’
LPS
///,// k\\\\\\ ,,//,, .\.\\\\ ,,,,// \\\\\NX ,//,/, k\\\\\\ ,,,/// .\\\\\\ - -----*
Shock
’
+ LPS
FIG. 4. Permeability index (1251-albumin alveolar lavage-to-plasma ratio) for shock (n = 3), LPS (n = 4), shock + LPS (n = 3), and normal animals (control, n = 4). Animals were killed 1 h after injection of 1251-albumin and alveolar lavage was performed. Values are means t SE. Shock + LPS group was statistically different from all other groups, but the other groups were not different from each other by analysis of variance and Scheffe’s F test (P < 0.05).
cannot be attributed to an inability to recognize or kill bacteria, because LPS was used as the second stimulus. Similar findings have been reported after burn injury. Burns are associated with an increased clinical incidence of MOFS, and experimental burns have been shown to increase the sensitivity to LPS (14). This increased sensitivity to LPS might be important in the pathogenesis of MOFS. The precise mechanism(s) that lead to the deleterious effects of LPS are not completely defined. LPS does not cause a direct toxic effect, at least at low concentrations, but the host immunoinflammatory system might be overstimulated, resulting in injury to the host as a result of LPS (10). Mononuclear phagocytes and the vascular endothelium may be responsible for the host response to LPS. In vitro experiments have shown that LPS causes activation of mononuclear phagocytes, leading to production of cytokines such as tumor necrosis factor (TNF) and interleukin-1. These cytokines cause activation of endothelial cells, leading to expression of a proinflammatory and procoagulant phenotype. In addition, the in vivo deleterious effects of LPS are mimicked at least in part by TNF, and the effects of LPS can be reduced by administration of anti-TNF antibody (2). Thus it is possible that part of the interaction of hemorrhagic shock and LPS administration may result from stimulation or “priming” of monocytes as a result of hemorrhagic shock. The pathophysiological changes that occur after hemorrhagic shock leading to the increased sensitivity to intravenously administered LPS are undetermined. Several possibilities present themselves. Hemorrhagic shock may lead to stimulation and/or priming of the host macrophage population, resulting in increased production of cytokines, both constitutively and in response to LPS. This possibility is supported by the recent demonstration of enhanced Kuppfer cell production of interleukin-1 and TNF in the first 24 h after shock (1). Neutrophils have been implicated as important media-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.
SENSITIVITY
TO
ENDOTOXIN
tors of organ injury in both shock and sepsis, and their adherence is thought to be critical in causing an injury. The profound and prolonged leukopenia after LPS administration in animals subjected to shock and LPS may be the result of proadhesive alterations resulting from cytokines released during or shortly after shock. The endothelium is known to become proadhesive as a result of surface expression of adhesion molecules after stimulation with cytokines (3, 9). Increased endothelial adherence molecule expression in response to sepsis but not hemorrhagic shock has been reported (16). Thus, if the number of adhesion molecules on the endothelial surface is increased by LPS after hemorrhage compared with no hemorrhage, it would be the result of priming by hemorrhage rather than direct stimulation. Hemorrhagic shock has also been associated with gastrointestinal tract injury and subsequent breakdown of the gastrointestinal mucosal barrier, allowing translocation of bacteria and bacterial products (4,17). Thus hemorrhage may cause a transient bacteremia or endotoxemia, leading to a generalized Shwartzman reaction. The generalized Shwartzman reaction is produced by intravenous injection of LPS followed by a second intravenous dose 18-24 h later. It has remained an immunopathological curiosity for over 60 years, and it is possible that hemorrhage followed by sepsis is the clinical analogue of a generalized Shwartzmann reaction. We conclude that hemorrhagic shock and resuscitation result in a proinflammatory state by 24 h after shock rather than a suppressed state, and this leads to increased sensitivity to the subsequent exposure to LPS. Understanding this increased sensitivity to LPS may contribute to a better understanding of the pathogenesis of MOFS and allow for development of therapeutic drugs for the treatment of this life-threatening syndrome. This study was supported by National Institutes of Health Grants HL-43141, HL-30542, GM-07037, and GM-42686 and by the American Heart Association of Washington. This study was approved by the Animal Care Committee of the University of Washington, and all experiments were performed in accordance with National Institutes of Health guidelines for the use of laboratory animals. Address for reprint requests: R. K. Winn, Dept. of Surgery ZA-16, Harborview Medical Center, 325 9th Ave., Seattle, WA 98104. Received
17 December
1990; accepted
in final
form
21 April
1992.
REFERENCES 1. AYALA, A., M. M. PERRIN, AND I. H. CHAUDRY. Increased susceptibility to sepsis following hemorrhage: defective Kupffer cell-mediated antigen presentation. Surg. Forum 40: 102-104, 1989. 2. BEUTLER, B., I. W. MISLARK, AND A. C. CERAMI. Passive immunization against cachectin/tumor necrosis factor protects mice from lethal effect of endotoxin. Science Wash. DC 234: 370-373, 1985. 3. BEVILACQUA, M., J. POBER, M. WHEELER, R. COTRAN, AND M.
AFTER
HEMORRHAGIC
SHOCK
1149
GIMBRONE. Interleukin 1 acts on cultured human vascular endothelium to increase the adhesion of polymorphonuclear leukocytes, monocytes, and related leukocyte cell lines. J. CLin. Inuest. 76: 2003-2010, 1985. 4. CAMERINI, D., S. JAMES, I. STAMENKOVICH, AND B. SEED. Leu-81 TQl is a human equivalent of MEL-14 lymph node homing receptor. Nature Lond. 342: 78-82, 1989. 5. CARRICO, C. J., J. L. MEAKINS, J. C. MARSHALL, D. FRY, AND R. V. MAIER. Multiple organ failure syndrome. Arch. Surg. 121: 196-208, 1986. 6. ESRIG, B. C., L. FRAZEE, S. F. STEPHENSON, H. C. POLK, R. L. FULTON, AND C. E. JONES. The predisposition to infection following hemorrhagic shock. Surg. Gynecol. Obstet. 144: 915-917, 1977. 7. FAIST, E., A. E. BAUE, H. DITTMER, AND G. HEBERER. Multiple organ failure in polytrauma patients. J. Trauma 23: 775-787, 1983. 8. FAIST, E., A. MEWES, T. STRASSER, S. WALZ, C. BAKER, W. ERTEL, AND G. HBERER. Alteration of monocyte function following major injury. Arch. Surg. 123: 287-292, 1988. 9. GAMBLE, J., J. HARLAN, S. KLEBANOFF, AND M. VADAS. Stimulation of the adherence of neutrophils to umbilical vein endothelium by human recombinant tumor necrosis factor. Proc. Natl. Acad. Sci. USA 82: 8667-8671, 1985. 10. LARRICK, J. W. Antibody inhibition of the immunoinflammatory cascade. J. Crit. Care 4: 211-224, 1989. 11. MEAKINS, J. L. Etiology of multiple organ failure. J. Trauma 30: S165S168, 1990. 12. MILESKI, W. J., R. K. WINN, N. V. VEDDER, T. H. POHLMAN, J. M. HARLAN, AND C. L. RICE. Inhibition of CD18-dependent neutrophil adherence reduces organ injury after hemorrhagic shock in primates. Surgery 108: 205-212, 1990. 13. MORRISON, D. C. Bacterial endotoxins and pathogenesis. Rev. Infect. Dis. 5: S733-S747, 1983. 14. NERLICH, M., J. FLYNN, AND R. H. DEMLING. Effect of thermal injury on endotoxin-induced lung injury. Surgery 93: 289-296, 1983. 15. POLK, H. C. The enhancement of host defenses against infection. Search for the holy grail? Surgery 99: l-6, 1986. 16. REDL, H., G. SCHLAG, H. P. DINGES, W. BURRMAN, R. ROTHLEIN, J. POBER, AND R. COTRAN. Endothelial activation after polytrauma and sepsis in the baboon (Abstract). Circ. Shock 31: 79, 1990. 17. SORI, A. J., B. F. RUSH, T. W. LYSZ, S. SMITH, AND G. W. MACHIEDO. The gut as source of sepsis after hemorrhagic shock. Am. J. Surg. 155: 187-192, 1988. 18. SIJFFREDINI, A. F., R. E. FROMM, M. M. PARKER, M. BRENNER, J. A. KOVACS, R. A. WESLEY, AND J. E. PARRILLO. The cardiovascular response of normal humans to the administration of endotoxin. N. Engl. J. Med. 321: 280-287, 1989. 19. VANDEVTNER, S. J. H., H. R. BULLER, J. W. CATE, A. STURK, AND W. PAUW. Endotoxemia: an early predictor of septicemia in febrile patients. Lancet 1: 605-609, 1988. 20. VEDDER, N. B., B. W. FOUTY, R. K. WINN, J. M. HARLAN, AND C. L. RICE. Role of neutrophils in generalized reperfusion injury associated with resuscitation from shock. Surgery 106: 509-516, 1989. 21. VEDDER, N. B., R. K. WINN, C. L. RICE, E. CHI, K.-E. ARFORS, AND J. M. HARLAN. A monoclonal antibody to the adherence promoting leukocyte glycoprotein CD18 reduces organ injury and improves survival from hemorrhagic shock and resuscitation in rabbits. J. Clin. Inuest. 81: 939-944, 1988. 22. WORTHEN, G. S., C. HASLETT, A. J. REES, R. S. GUMBAY, J. E. HENSON, AND P. M. HENSON. Synergistic effect of trace amounts of lipopolysaccharide and neutrophil stimuli on vascular permeability and neutrophil sequestration in the lung. Am. Rev. Respir. Dis. 136: 19-28, 1987.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.
