- Email: [email protected]
A New Approach to the Treatment of Experimental Septic Shock
JOURNAL OF SURGICAL RESEARCH ARTICLE NO.
61, 311–316 (1996)
0122
A New Approach to the Treatment of Experimental Septic Shock ROBERT M. HARDAWAY, M.D.,1 CHARLES H. WILLIAMS, PH.D.,
AND
YANG SUN, M.D.
Departments of Surgery and Anesthesiology, Texas Tech University School of Medicine, El Paso, Texas 79905 Submitted for publication June 13, 1995
Previous work has shown that disseminated intravascular coagulation (DIC) may produce multiple organ failure, including adult respiratory distress syndrome, by obstruction of visceral micro circulation by microclots. DIC can be produced by sepsis. This study tests the ability of a plasminogen activator to prevent death from an intravenous injection of killed Escherichia coli by causing lysis of the microclots. Subjects were two groups of 8 pigs each with body weight of 60–70 lbs. Killed Escherichia coli were injected IV in 16 pigs. Invasive monitoring was used to record physiologic data during the 5.0-hr experimental period. Urokinase injected 20 min after the injection of Escherichia coli organisms significantly prevented mortality, acidosis, and development of blood incoagulability. We conclude that plasminogen activator can significantly prevent fatal Escherichia coli (septic) shock without causing bleeding. q 1996 Academic Press, Inc.
INTRODUCTION
This study tests the theory that Escherichia coli shock is caused by disseminated intravascular coagulation (DIC) and the production of microclots of the viscera. Can these microclots be lysed by the administration of a plasminogena activator? DIC was first described by Landois [1] in 1875 when he gave dogs human blood intravenously and found hyalin thrombi in the vessels of the mesentery. In 1951, Schneider noted a decrease in fibrinogen and microscopic fibrin embolism in a patient with placenta abruptio [2]. In 1954, Hardaway reported four cases of acute renal failure after trauma or sepsis, all of which showed decreased or absent fibrinogen, a clotting defect, and fibrin thrombi [3]. These cases were the first reported cases of DIC causing organ failure in humans. It was postulated that the clotting defect was due to the using up of clotting factors in an episode of disseminated intravascular coagulation and that the renal failure was due to fibrin microclots in capillaries of the kidney. After experimental work on dogs using admin1 To whom correspondence and reprint requests should be addressed at Texas Tech University School of Medicine, 4800 Alberta Avenue, El Paso, Texas 79905. Fax: (915) 545-6521.
istration of human blood, it was shown that DIC could produce acute renal failure [4], hemorrhagic necrosis in the stomach and intestine [5], liver failure [6], pancreatitis [7], and pulmonary failure [8]. In fact DIC could produce multiple organ failure. These syndromes were called ‘‘Syndromes of Disseminated Intravascular Coagulation’’ in an article published in 1961 [9]. This was expanded into a book on DIC in 1966 [10]. A clinical study of patients in severe shock due to trauma or sepsis or both confirmed DIC in nearly every case [11]. It was shown that endotoxin could cause intravascular clotting and could cause hemorrhagic lesions in the lungs, liver, kidney, and gut [12]. It was also shown that trauma to muscle could produce DIC and lesions in the lungs, liver, kidney, and gut [13, 14]. DIC could thus be initiated by trauma or sepsis or both. If DIC was the cause of organ failure, perhaps heparin would prevent the development of DIC. This was tried and found to help [15], but heparin had to be given before endotoxin was given. It was not the answer to DIC because: (1) Heparin is inactivated in a pH below 7.2 [16] and the pH of capillary blood in shock is usually below 7.2, (2) Heparin will not dissolve a clot that is already formed, (3) Heparin may initiate platelet agglutination, and (4) Heparin may cause hemorrhage. A better approach to the treatment of DIC is by clot dissolution. This was tried with good results in 1961, using a drug then available: fibrinolysin (an activated plasminogen) [17, 18]. However, the manufacture of this drug was discontinued. In recent years plasminogen activators have been introduced and found to be very effective in dissolution of large clots as in coronary infarction, pulmonary infarction, venous thrombosis, thrombotic stroke, and other conditions [19, 20]. Fibrinolysin produced dramatic results in traumatic and hemorrhagic shock in dogs [17, 18]. Later it was shown that by using urokinase or tissue plasminogen activator (tPA) it was possible to prevent and treat the syndrome of adult respiratory distress syndrome caused by trauma in pigs [21]. Urokinase would also prevent multiple organ failure in the kidneys and liver in traumatic shock in pigs, if given within 4 hr after trauma [22]. The most obvious cause of death in traumatic shock was acute pulmonary failure caused by massive involvement of the lungs by lesions of congestion, hemorrhage, edema, and microcirculatory obstruction, 0022-4804/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
311
AID
JSR 4701
/
m4999$$123
02-20-96 13:48:18
srga
AP: Surg Res
312
JOURNAL OF SURGICAL RESEARCH: VOL. 61, NO. 2, MARCH 1996
which was greatly decreased by urokinase. However, the administration of plasminogen activator for prevention and treatment of DIC due to Escherichia coli has never been tried. A computer search has not found any reference to this in the past 10 years. The present experiment tests the ability of a plasminogen activator to prevent death from an injection of killed Escherichia coli organisms. MATERIALS AND METHODS Sixteen healthy pigs weighing 60 to 70 lbs were used in the study. They were divided into two groups, control and treated. Neutrality in group selection was assured by scheduling the group assignment before the pig was seen. Control and treated pigs were alternated and scheduled before a pig was brought in from the farm. The study was approved by the Animal Use Committee. Pigs were anesthetized with pentothal, and an endotrachael tracheal tube was inserted and connected to a respirator giving 50% O2 . Two catheters were placed in a jugular vein. One was a Swan-Ganz Catheter and another was used for IV injection. A catheter was also placed in a carotid artery. A catheter was placed in the urinary bladder. A rectal probe measured rectal temperature. Electrocardiogram was monitored. Blood volume, central venous pressure (CVP), and some nutrition were maintained by an IV drip of 0.9% sodium chloride supplement with IV glucose. Continuous recordings were made on arterial pressure, pulmonary artery pressure, CVP, and blood temperature. Frequent measurements of cardiac output and pulmonary capillary wedge pressure were made. Blood samples were taken before injection of killed E. coli organisms, immediately after injection, at 1 hr, at 5 hr, and at 24 hr if the pig was still alive. Blood was analyzed for pO2 , pCO2 , pH in both arterial and venous blood, clotting time, fibrinogen, prothrombin time, partial thromboplastin time, fibrin split products, WBC, differential, platelets, urea nitrogen, creatinine, glutamic oxaloacetic transaminase, and glutamic pyruvic transaminase. The following hemodynamic parameters were derived using these standards equations: Mean arterial pressure (mAP) Å (systolic AP 0 diastolic AP)/3 / diastolic AP Cardiac index (CT) Å CO/kg body weight Stroke volume (SV) Å CO/HR Stroke volume index (SVI) Å SV/kg body weight (We use body weight rather than surface area for pig work) Rate pressure product (RPP) Å HR 1 systolic AP
FIG. 1. Pulmonary artery pressure in both groups of pigs.
corded by using a spectrophotometer. Thus, a standard line was established. A second culture was washed and centrifuged in the same way and adjusted to the concentration of 2 1 10/10/ml spectrophotometrically. Bacteria were killed by boiling for 30 min and stored at 707C until used in experiments. Before the storage, a subculture was made to determine if bacteria had been killed. All pigs were injected with 2 ml killed Escherichia coli organisms in 20 ml saline over a 20-min period, using a Sage mechanical syringe. Eight of the 16 pigs were given an injection of 250,000 U of Urokinase (Abbokinase, Abbott Laboratories, Chicago, IL) diluted in 20 ml saline and injected over a 20-min period, using the same mechanical syringe. This dosage of Urokinase was suggested by Abbott Laboratories. Injection of the urokinase was started 20 min after the Escherichia coli injection was completed. Control pigs were given a similar injection of normal saline without Urokinase. Anesthesia and catheters were continued for 6 hr, when blood samples were taken and catheters and anesthesia discontinued. A heparin lock was placed in each catheter for later blood sampling. No heparin entered the circulation. The animals were placed in a cage for observation and allowed to recover spontaneously. If alive in 24 hr, the animals were counted as survivors, blood samples were taken, and the animals were killed by lethal injection of 10 ml of 1 M KCl. All animals were autopsied for gross and microscopic pathology. The experiment was terminated after 24 hr because of the need for continuous monitoring and because most endotoxin experiments have used this as an end point. Statistics were run on an IBM compatible 386-33Mz computer.
Pulmonary vascular resistance (PVR) Å (mPAP 0 mCWP) 1 (80)/ CO
RESULTS
Total peripheral resistance (TPR) Å (mAP 0 CVP) 1 (80)/CO Right ventricular stroke work index (RVSWI) Å (mPAP 0 CVP) 1 .0136 Left ventricular stroke work index (LVSWI) Å (MAP 0 PCWP) 1 .0136 Coronary perfusion pressure (CPP) Å diastolic AP 0 PCWP Cardiac output was determined by thermodilution at 1-hr intervals. Ten cubic centimeters of 5% dextrose solution at 47C was injected via a central venous line. This procedure was performed in triplicate. A killed Escherichia coli culture was used in preference to purified edotoxin, because it is much more potent and much easier to calibrate a suitable fatal dose. A killed Escherichia coli culture was prepared as follows: Escherichia coli (strain 078:H11) from American Type Culture Collection (12301 Parklawn Drive, Rockville, Maryland 20852) was inoculated at 357C to 377C overnight (18–24 hr) in a shaking water bath. The cultures were centrifuged and washed three times in pyrogen-free saline, and were made into series dilution. The individual solution was cultured again by using a calibrated loop for determining bacterial concentration/ml, and the value of OD re-
AID
JSR 4701
/
m4999$$124
02-20-96 13:48:18
All eight pigs in the control and treated groups survived for the first 5 hr of the experiment. All eight treated pigs survived for the next 20 hr. Four untreated control pigs died before 12 hr had elapsed and seven had died before 24 hr. Only one survived 24 hr. All animals alive at 24 hr were killed and autopsied. The treated pigs still alive seemed in good shape and could possibly have survived permanently. Difference in mortality was significant at P õ 0.02. At autopsy there were few abnormal findings in tissue except the lungs. Control pigs’ lungs were atelectatic, congested, cyanotic, and edematous. Treated pigs’ lungs were pink and aerated. Pulmonary artery pressure always rose markedly about 10 min after the start of the 20-min injection of killed Escherichia coli (Fig. 1). In control pigs this high pulmonary artery pressure was sustained for 1 hr or more; however, in treated pigs, the
srga
AP: Surg Res
HARDAWAY, WILLIAMS, AND SUN: TREATMENT OF Escherichia coli
313
FIG. 2. Cardiac output in both groups of pigs.
pressure returned to normal about 10 min after start of the 20-min injection of Urokinase. Cardiac output (Fig. 2) increased in control animals but decreased in treated animals. Core temperature started at 37.87C and significantly increased (P Å 0.00232) to 39.97C at 5 hr. There were no significant differences in core temperatures between control and treated animals. There was a significant decrease in arterial hydrogen ion concentration (P Å 0.0221 E-04) in the control group of animals, (Fig. 3). Arterial PO2 levels decreased significantly in both groups (Fig. 4). The 24-hr PO2 for controls is skewed as it represents only the one surviving pig. Arterial base excess was significantly (P Å 0.00345) decreased more markedly in the control group (Fig. 5). There was a significant decrease of fibrinogen (Fig. 6) in the control animals, reflecting a using up of fibrinogen in DIC. There was a significant rise in fibrinogen in treated animals. (Fibrinogen manufacture is rapid as a result of any stress.) White blood cells (Fig. 7) rose significantly in the control pigs but remained relatively consistent in the treated pigs. Fibrin split products increased significantly in all pigs, denoting production of DIC. Clotting time remained normal in all pigs except for the one surviving control pig, when it failed to clot in two hours. (Fig. 8) Urinary output increased in all pigs (Fig. 9).
FIG. 4. Arterial pO2 in both groups of pigs.
DISCUSSION
Many septic shock models use live E. coli organisms. Therefore this model may not be as clinically relevant as they. However, this model of ‘‘septic’’ shock was chosen because previous work had shown that killed Escherichia coli organisms were much more lethal than ‘‘endotoxin’’ [43]. It was postulated that the major cause of the DIC in septic shock was the thrombogenic quality of phospholipids, which line the inner surface of all cell membranes including bacteria [44]. If bacteria are killed by antibodies or by antibiotics, this thrombogenic phospholipid is exposed to the circulating blood and may initiate DIC. Elaine Tuomanen in an article in Scientific American [32] tells how her friend, Alexander Tomay, ‘‘accidentally challenged animals with killed pneumococci and yet the animals got sick. Dead bacteria seemed as harmful as living ones. How could this be?’’ [45]. The Herxheimer reaction may be due to the broken cell walls of the spirochete. Ascites fluid contains many broken cells which, when reinjected into the circulation by the Leveen shunt, results in DIC. If one injects human amniotic fluid (which contains many broken cells) intravenously into a dog, DIC results [46]. The pulmonary artery pressure rises, blood may become incoagulable, and multiple organ failure may result [33]. If one
FIG. 3. Arterial hydrogen ion concentration.
AID
JSR 4701
/
m4999$$124
02-20-96 13:48:18
FIG. 5. Arterial base excess in both groups of pigs.
srga
AP: Surg Res
314
JOURNAL OF SURGICAL RESEARCH: VOL. 61, NO. 2, MARCH 1996
FIG. 8. Clotting times in both groups of pigs.
FIG. 6. Fibrinogen levels in both groups of pigs.
filters the amniotic fluid before injecting it, the fluid is harmless [33]. It was further postulated that the mechanism of multiple organ failure in septic shock was the obstruction of visceral capillaries by microclots of DIC [3]. If this were true, the dissolution of the microclots would restore circulation to the organs. In this experiment, the administration of plasminogen activator resulted in a significantly decreased mortality and improvement in a number of other parameters, especially pH. It is difficult to prove that DIC actually occurs. Fibrin clots are sometimes found in autopsy material, but often are not. Many of the control pigs showed microscopic fibrin thrombi, as has been seen in other experiments [26]. It is theorized that the body’s fibrinolytic mechanism dissolves many of the microclots immediately after death. Perhaps the best indicator of DIC is the elevation of fibrin split products which occurred in all pigs. A theoretical danger of administering a plasminogen activator in the presence of shock is the development of a clotting defect and bleeding. This did not occur. In fact, the only clotting time increase was in a control pig. Clotting parameters otherwise
remained normal. In fact, fibrinogen was actually higher in the treated animals. Fibrinogen has been shown to increase very rapidly in response to any stress [23]. The fibrinogen level in the blood is determined by the balance between fibrinogen production and conversion to fibrin. In spite of using up of the fibrinogen by the DIC process, it has been found that in most cases of clinical DIC, fibrinogen is high [24, 25]. Many drugs have been tried to treat septic shock. Of course antibiotics, respiratory support, metabolic support by fluids and electrolytes, and nutritional support are all essential and universally used. However, mortality from septic shock, particularly due to multiple organ failure, especially adult respiratory distress syndrome, remains very high (50 to 90%) [27–34]. Corticosteroids have been extensively used with generally poor results [34, 35]. Vasoactive drugs, including vasoconstrictors, vasodilators, and mixed types have been helpful [36]. Endogenous opiates have been used with mixed results [37]. Many centers have been using monoclonal antibodies with great success in the laboratory [38–40] and also in some clinical studies [39, 40]. However, in an FDA-approved multicenter study, many unexpected deaths occurred which resulted in the removal from the market of interleukin-1 receptor antagonist [39]. A recent study of monoclonal antibody treatment showed an actual increase in mortality in canine gram-negative septic shock [41]. None of the above drug treatments have been shown to alter the high mortality in septic shock. Death from septic shock in the United States increased from 35,000 in 1970 to
FIG. 7. WBC in both groups of pigs.
AID
JSR 4701
/
m4999$$124
02-20-96 13:48:18
FIG. 9. Urinary output.
srga
AP: Surg Res
HARDAWAY, WILLIAMS, AND SUN: TREATMENT OF Escherichia coli
94,000 in 1980 to 198,000 in 1990 [42]. The death rate from septic shock per 100,000 population increased from 1.7 in 1970 to 4.2 in 1980 to 7.9 in 1990, according to the Statistical Abstract of the United States, 1992 [42]. For years investigators have focused on the pivotal role of endotoxin. It is composed of an O-polysaccharide chain, a core sugar, and a lipophilic fatty acid. The lipid a portion of the lipopolysaccharide includes expression of cytokine genes through stimulation of receptors on the surface of target cells and activation of the nuclear transcription factor NF-KB. This gene activation brings on the inflammatory response and its attendant deliterius effects, i.e., hypotension and organ disfunction. Because it causes a shock-like state when injected into experimental animals, endotoxin has been implicated as a causative agent in septic shock. However, as the understanding of sepsis is now more detailed, it is clear that endotoxin is only one component of sepsis. This may explain why trials in which antiendotoxin antibodies were injected into patients have failed to show consistent benefit.
9. 10.
11.
12.
13. 14. 15.
16. 17. 18.
CONCLUSION
Killed Escherichia coli organisms given intravenously cause death, with a low arterial and venous pH, high cardiac output, increased core temperature, elevated white cell count, and other metabolic changes. Death and the above changes in parameters can be significantly prevented by a plasminogen activator. Administration of a plasminogen activator in the presence of Escherichia coli shock causes no changes in the clotting mechanisms; in fact, it seems to prevent these changes. These results are consistent with the theory that DIC is produced by Escherichia coli organisms, that DIC causes the production of microclots which occlude the microcirculation of the viscera, and that a plasminogen activator will cause dissolution of these microclots with improvement in the circulation to the viscera.
19.
20. 21.
22.
23. 24.
25. 26.
REFERENCES
27.
1. Landois Transfusion des Blutes, Lypzig, 1875. 2. Schneider, C. L. Fibrin embolism with defibrillation as one of the end results during placenta abruptio. SGO 92: 1951. 3. Hardaway, R. M., Mckay, D. G., and Williams, J. H. Lower nephron nephrosis. Am. J. Surg. 87: 41–49, 1954. 4. Hardaway, R. M., and McKay, D. G. Production of lower nephron nephrosis in dogs by means of an episode of intravascular clotting. Surg. Form. IX: 134–137, 1959. 5. Hardaway, R. M., and McKay, D. G. Intravascular thrombi and the intestinal factor of irreversible shock. Ann. Surg. 150: 261– 265, 1959. 6. Hardaway, R. M., and McKay, D. G. Pathologic study of intravascular coagulation following incompatible blood transfusion in dogs. Am. J. Surg. 91: 24–40, 1956. 7. Hardaway, R. M., and McKay, D. G. Production of acute hemorrhagic pancreatitis in dogs by means of an episode of DIC. Surgery 45: 557–561, 1959. 8. Hardaway, R. M. Pulmonary microthrombi. In A. A. Sasahara
AID
JSR 4701
/
m4999$$124
02-20-96 13:48:18
28.
29.
30. 31.
32.
33.
34.
315
(Ed.), Pulmonary Embolic Disease, pp. 65–72. New York: Grune-Stratton, 1965. Hardaway, R. M. Disseminated intravascular coagulation syndromes. Arch. Surg. 83: 842–850, 1961. Hardaway, R. M. Syndromes of Disseminated Intravascular Coagulation with Special Reference to Shock and Hemorrhage. Springfield, IL: Charles C Thomas Pub., 1966. Hardaway, R. M., James, J. M., and Anderson, R. W., et al. Intensive study and treatment of shock in man. JAMA 199: 779–790, 1967. Hardaway, R. M., Husni, E. A., and Geever, E. F., et al. Endotoxin shock, a manifestation of intravascular coagulation. Ann. Surg. 154: 791–802, 1961. Hardaway, R. M. Mechanism of traumatic shock. SGO 151: 65– 69, 1980. Hardaway, R. M., and Williams, C. H. A new treatment of traumatic shock and ARDS. Resuscitation 19: 61–75, 1990. Hardaway, R. M., Hasni, E. A., Eseever, E. F., et al. Endotoxin shock, a manifestation of intravascular coagulation. Ann. Surg. 154: 791–802, 1961. Hardaway, R. M., Clovitz, H. O., Brewster, W. R., et al. Clotting time of heparanized blood. Arch. Surg. 89: 701–705, 1964. Hardaway, R. M., and Johnson, D. G. Influence of fibrinolysis on shock. JAMA 183: 397–599, 1963. Hardaway, R. M., and Burns, V. W. Mechanism of action of fibrinolysis in the prevention of irreversible hemorrhagic shock. Ann. Surg. 157: 305–309, 1963. Wall, T. C., et al. Results of high dose intravenous urokinase for acute myocardial infarction. Am. J. Cardiol. 65(3): 124–131, 1990. Nadrowski, L. F. Deep venous thrombosis: Recent advances in pathogenesis and treatment. Surg. Ann. 23(1): 1991. Hardaway, R. M., Williams, C. H., and Marvasti, M., et al. Prevention of adult respiratory distress syndrome with plasminogen activator in pigs. Crit. Care Med. 18: 1413–1418, 1990. Hardaway, R. M., and Williams, C. H. Prevention of multiple organ failure with plasminogen activator. Current Ther. Res. 49: 711–722, 1991. Hardaway, R. M., Johnson, O. G., et al. Influence of stress on fibrinogen. J. Trauma 4: 673–676, 1964. Hardaway, R. M., James, P. M., Anderson, R. W., et al. Intensive study and treatment of shock in man. JAMA 199: 779–790, 1967. Hardaway, R. M. The significance of coagulation and thrombotic changes after injury. J. Trauma 354–357, 1969. Hardaway, R. M. Syndromes of Disseminated Intravascular Coagulation. Springfield, IL: Charles C Thomas, Pub. 1966. Hardaway, R. M. Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. Hardaway, R. M. Treatment of adult respiratory distress syndrome with plasminogen activator. Hematol. Rev. Commun., in press. Hardaway, R. M., and Williams, C. H. Prevention of multiple organ failure with plasminogen activator. Curr. Ther. Res. 49: 711–722, 1991. Hardaway, R. M., and Johnson, D. Clotting mechanism in endotoxic block. Arch. Int. Med. 112: 775–782, 1963. Hinshaw, L. B., and Beller, B. K. Septic shock, physiological aspects. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publisher, 1988. Archer, L. T. Septic shock, pathological aspects. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. Schumer, W. Septic shock—Clinical aspects. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. Schumer, W. Treatment of septic shock with steroids. In R. M.
srga
AP: Surg Res
316
JOURNAL OF SURGICAL RESEARCH: VOL. 61, NO. 2, MARCH 1996 Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988.
35. Shatney, C. H. The role of corticosteroids in the treatment of shock. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. 36. MacConnell, K. L. Vasoactive drugs in the treatment of shock. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. 37. Bowen, J. C., and Rees, M. Endogenous opiates in shock. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. 38. Gaffin, S. L. Antibody therapy in shock. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. 39. Dudley, M. N. Interleukin-1 receptor antagonists significantly
AID
JSR 4701
/
m4999$$124
02-20-96 13:48:18
40. 41.
42. 43. 44. 45. 46.
reduce death from sepsis, University of Rhode Island MultiCenter Study, Gen. Surg. News Feb.: 28–29, 1993. Editor’s Note, Gen. Surg. News Feb.: 28, 1993. Quezado, Z. M. N., Natnason, C., Assinf, D. W., et al. A controlled trial of HA-IA in a canine model of gram-negative septic shock. JAMA 269: 2221–2227, 1993. Statistical Abstract of the United States, 1992. Hardaway, R. M. ‘‘Trauma, sepsis and disseminated intravascular coagulation.’’ J. Intensive Care Med., in press. Hardaway, R. M. Broken cell walls. The ‘‘toxin’’ of traumatic and septic shock. Critical Care Int., in press. Tuomanen, E. Breaching the blood-brain barrier. Sci. Am. Feb.: 80–84, 1993. Hardaway, R. M., and McKay, D. G. Disseminated intravascular coagulation: A cause of shock. Ann. Surg. 149: 462–470, 1959.
srga
AP: Surg Res
61, 311–316 (1996)
0122
A New Approach to the Treatment of Experimental Septic Shock ROBERT M. HARDAWAY, M.D.,1 CHARLES H. WILLIAMS, PH.D.,
AND
YANG SUN, M.D.
Departments of Surgery and Anesthesiology, Texas Tech University School of Medicine, El Paso, Texas 79905 Submitted for publication June 13, 1995
Previous work has shown that disseminated intravascular coagulation (DIC) may produce multiple organ failure, including adult respiratory distress syndrome, by obstruction of visceral micro circulation by microclots. DIC can be produced by sepsis. This study tests the ability of a plasminogen activator to prevent death from an intravenous injection of killed Escherichia coli by causing lysis of the microclots. Subjects were two groups of 8 pigs each with body weight of 60–70 lbs. Killed Escherichia coli were injected IV in 16 pigs. Invasive monitoring was used to record physiologic data during the 5.0-hr experimental period. Urokinase injected 20 min after the injection of Escherichia coli organisms significantly prevented mortality, acidosis, and development of blood incoagulability. We conclude that plasminogen activator can significantly prevent fatal Escherichia coli (septic) shock without causing bleeding. q 1996 Academic Press, Inc.
INTRODUCTION
This study tests the theory that Escherichia coli shock is caused by disseminated intravascular coagulation (DIC) and the production of microclots of the viscera. Can these microclots be lysed by the administration of a plasminogena activator? DIC was first described by Landois [1] in 1875 when he gave dogs human blood intravenously and found hyalin thrombi in the vessels of the mesentery. In 1951, Schneider noted a decrease in fibrinogen and microscopic fibrin embolism in a patient with placenta abruptio [2]. In 1954, Hardaway reported four cases of acute renal failure after trauma or sepsis, all of which showed decreased or absent fibrinogen, a clotting defect, and fibrin thrombi [3]. These cases were the first reported cases of DIC causing organ failure in humans. It was postulated that the clotting defect was due to the using up of clotting factors in an episode of disseminated intravascular coagulation and that the renal failure was due to fibrin microclots in capillaries of the kidney. After experimental work on dogs using admin1 To whom correspondence and reprint requests should be addressed at Texas Tech University School of Medicine, 4800 Alberta Avenue, El Paso, Texas 79905. Fax: (915) 545-6521.
istration of human blood, it was shown that DIC could produce acute renal failure [4], hemorrhagic necrosis in the stomach and intestine [5], liver failure [6], pancreatitis [7], and pulmonary failure [8]. In fact DIC could produce multiple organ failure. These syndromes were called ‘‘Syndromes of Disseminated Intravascular Coagulation’’ in an article published in 1961 [9]. This was expanded into a book on DIC in 1966 [10]. A clinical study of patients in severe shock due to trauma or sepsis or both confirmed DIC in nearly every case [11]. It was shown that endotoxin could cause intravascular clotting and could cause hemorrhagic lesions in the lungs, liver, kidney, and gut [12]. It was also shown that trauma to muscle could produce DIC and lesions in the lungs, liver, kidney, and gut [13, 14]. DIC could thus be initiated by trauma or sepsis or both. If DIC was the cause of organ failure, perhaps heparin would prevent the development of DIC. This was tried and found to help [15], but heparin had to be given before endotoxin was given. It was not the answer to DIC because: (1) Heparin is inactivated in a pH below 7.2 [16] and the pH of capillary blood in shock is usually below 7.2, (2) Heparin will not dissolve a clot that is already formed, (3) Heparin may initiate platelet agglutination, and (4) Heparin may cause hemorrhage. A better approach to the treatment of DIC is by clot dissolution. This was tried with good results in 1961, using a drug then available: fibrinolysin (an activated plasminogen) [17, 18]. However, the manufacture of this drug was discontinued. In recent years plasminogen activators have been introduced and found to be very effective in dissolution of large clots as in coronary infarction, pulmonary infarction, venous thrombosis, thrombotic stroke, and other conditions [19, 20]. Fibrinolysin produced dramatic results in traumatic and hemorrhagic shock in dogs [17, 18]. Later it was shown that by using urokinase or tissue plasminogen activator (tPA) it was possible to prevent and treat the syndrome of adult respiratory distress syndrome caused by trauma in pigs [21]. Urokinase would also prevent multiple organ failure in the kidneys and liver in traumatic shock in pigs, if given within 4 hr after trauma [22]. The most obvious cause of death in traumatic shock was acute pulmonary failure caused by massive involvement of the lungs by lesions of congestion, hemorrhage, edema, and microcirculatory obstruction, 0022-4804/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
311
AID
JSR 4701
/
m4999$$123
02-20-96 13:48:18
srga
AP: Surg Res
312
JOURNAL OF SURGICAL RESEARCH: VOL. 61, NO. 2, MARCH 1996
which was greatly decreased by urokinase. However, the administration of plasminogen activator for prevention and treatment of DIC due to Escherichia coli has never been tried. A computer search has not found any reference to this in the past 10 years. The present experiment tests the ability of a plasminogen activator to prevent death from an injection of killed Escherichia coli organisms. MATERIALS AND METHODS Sixteen healthy pigs weighing 60 to 70 lbs were used in the study. They were divided into two groups, control and treated. Neutrality in group selection was assured by scheduling the group assignment before the pig was seen. Control and treated pigs were alternated and scheduled before a pig was brought in from the farm. The study was approved by the Animal Use Committee. Pigs were anesthetized with pentothal, and an endotrachael tracheal tube was inserted and connected to a respirator giving 50% O2 . Two catheters were placed in a jugular vein. One was a Swan-Ganz Catheter and another was used for IV injection. A catheter was also placed in a carotid artery. A catheter was placed in the urinary bladder. A rectal probe measured rectal temperature. Electrocardiogram was monitored. Blood volume, central venous pressure (CVP), and some nutrition were maintained by an IV drip of 0.9% sodium chloride supplement with IV glucose. Continuous recordings were made on arterial pressure, pulmonary artery pressure, CVP, and blood temperature. Frequent measurements of cardiac output and pulmonary capillary wedge pressure were made. Blood samples were taken before injection of killed E. coli organisms, immediately after injection, at 1 hr, at 5 hr, and at 24 hr if the pig was still alive. Blood was analyzed for pO2 , pCO2 , pH in both arterial and venous blood, clotting time, fibrinogen, prothrombin time, partial thromboplastin time, fibrin split products, WBC, differential, platelets, urea nitrogen, creatinine, glutamic oxaloacetic transaminase, and glutamic pyruvic transaminase. The following hemodynamic parameters were derived using these standards equations: Mean arterial pressure (mAP) Å (systolic AP 0 diastolic AP)/3 / diastolic AP Cardiac index (CT) Å CO/kg body weight Stroke volume (SV) Å CO/HR Stroke volume index (SVI) Å SV/kg body weight (We use body weight rather than surface area for pig work) Rate pressure product (RPP) Å HR 1 systolic AP
FIG. 1. Pulmonary artery pressure in both groups of pigs.
corded by using a spectrophotometer. Thus, a standard line was established. A second culture was washed and centrifuged in the same way and adjusted to the concentration of 2 1 10/10/ml spectrophotometrically. Bacteria were killed by boiling for 30 min and stored at 707C until used in experiments. Before the storage, a subculture was made to determine if bacteria had been killed. All pigs were injected with 2 ml killed Escherichia coli organisms in 20 ml saline over a 20-min period, using a Sage mechanical syringe. Eight of the 16 pigs were given an injection of 250,000 U of Urokinase (Abbokinase, Abbott Laboratories, Chicago, IL) diluted in 20 ml saline and injected over a 20-min period, using the same mechanical syringe. This dosage of Urokinase was suggested by Abbott Laboratories. Injection of the urokinase was started 20 min after the Escherichia coli injection was completed. Control pigs were given a similar injection of normal saline without Urokinase. Anesthesia and catheters were continued for 6 hr, when blood samples were taken and catheters and anesthesia discontinued. A heparin lock was placed in each catheter for later blood sampling. No heparin entered the circulation. The animals were placed in a cage for observation and allowed to recover spontaneously. If alive in 24 hr, the animals were counted as survivors, blood samples were taken, and the animals were killed by lethal injection of 10 ml of 1 M KCl. All animals were autopsied for gross and microscopic pathology. The experiment was terminated after 24 hr because of the need for continuous monitoring and because most endotoxin experiments have used this as an end point. Statistics were run on an IBM compatible 386-33Mz computer.
Pulmonary vascular resistance (PVR) Å (mPAP 0 mCWP) 1 (80)/ CO
RESULTS
Total peripheral resistance (TPR) Å (mAP 0 CVP) 1 (80)/CO Right ventricular stroke work index (RVSWI) Å (mPAP 0 CVP) 1 .0136 Left ventricular stroke work index (LVSWI) Å (MAP 0 PCWP) 1 .0136 Coronary perfusion pressure (CPP) Å diastolic AP 0 PCWP Cardiac output was determined by thermodilution at 1-hr intervals. Ten cubic centimeters of 5% dextrose solution at 47C was injected via a central venous line. This procedure was performed in triplicate. A killed Escherichia coli culture was used in preference to purified edotoxin, because it is much more potent and much easier to calibrate a suitable fatal dose. A killed Escherichia coli culture was prepared as follows: Escherichia coli (strain 078:H11) from American Type Culture Collection (12301 Parklawn Drive, Rockville, Maryland 20852) was inoculated at 357C to 377C overnight (18–24 hr) in a shaking water bath. The cultures were centrifuged and washed three times in pyrogen-free saline, and were made into series dilution. The individual solution was cultured again by using a calibrated loop for determining bacterial concentration/ml, and the value of OD re-
AID
JSR 4701
/
m4999$$124
02-20-96 13:48:18
All eight pigs in the control and treated groups survived for the first 5 hr of the experiment. All eight treated pigs survived for the next 20 hr. Four untreated control pigs died before 12 hr had elapsed and seven had died before 24 hr. Only one survived 24 hr. All animals alive at 24 hr were killed and autopsied. The treated pigs still alive seemed in good shape and could possibly have survived permanently. Difference in mortality was significant at P õ 0.02. At autopsy there were few abnormal findings in tissue except the lungs. Control pigs’ lungs were atelectatic, congested, cyanotic, and edematous. Treated pigs’ lungs were pink and aerated. Pulmonary artery pressure always rose markedly about 10 min after the start of the 20-min injection of killed Escherichia coli (Fig. 1). In control pigs this high pulmonary artery pressure was sustained for 1 hr or more; however, in treated pigs, the
srga
AP: Surg Res
HARDAWAY, WILLIAMS, AND SUN: TREATMENT OF Escherichia coli
313
FIG. 2. Cardiac output in both groups of pigs.
pressure returned to normal about 10 min after start of the 20-min injection of Urokinase. Cardiac output (Fig. 2) increased in control animals but decreased in treated animals. Core temperature started at 37.87C and significantly increased (P Å 0.00232) to 39.97C at 5 hr. There were no significant differences in core temperatures between control and treated animals. There was a significant decrease in arterial hydrogen ion concentration (P Å 0.0221 E-04) in the control group of animals, (Fig. 3). Arterial PO2 levels decreased significantly in both groups (Fig. 4). The 24-hr PO2 for controls is skewed as it represents only the one surviving pig. Arterial base excess was significantly (P Å 0.00345) decreased more markedly in the control group (Fig. 5). There was a significant decrease of fibrinogen (Fig. 6) in the control animals, reflecting a using up of fibrinogen in DIC. There was a significant rise in fibrinogen in treated animals. (Fibrinogen manufacture is rapid as a result of any stress.) White blood cells (Fig. 7) rose significantly in the control pigs but remained relatively consistent in the treated pigs. Fibrin split products increased significantly in all pigs, denoting production of DIC. Clotting time remained normal in all pigs except for the one surviving control pig, when it failed to clot in two hours. (Fig. 8) Urinary output increased in all pigs (Fig. 9).
FIG. 4. Arterial pO2 in both groups of pigs.
DISCUSSION
Many septic shock models use live E. coli organisms. Therefore this model may not be as clinically relevant as they. However, this model of ‘‘septic’’ shock was chosen because previous work had shown that killed Escherichia coli organisms were much more lethal than ‘‘endotoxin’’ [43]. It was postulated that the major cause of the DIC in septic shock was the thrombogenic quality of phospholipids, which line the inner surface of all cell membranes including bacteria [44]. If bacteria are killed by antibodies or by antibiotics, this thrombogenic phospholipid is exposed to the circulating blood and may initiate DIC. Elaine Tuomanen in an article in Scientific American [32] tells how her friend, Alexander Tomay, ‘‘accidentally challenged animals with killed pneumococci and yet the animals got sick. Dead bacteria seemed as harmful as living ones. How could this be?’’ [45]. The Herxheimer reaction may be due to the broken cell walls of the spirochete. Ascites fluid contains many broken cells which, when reinjected into the circulation by the Leveen shunt, results in DIC. If one injects human amniotic fluid (which contains many broken cells) intravenously into a dog, DIC results [46]. The pulmonary artery pressure rises, blood may become incoagulable, and multiple organ failure may result [33]. If one
FIG. 3. Arterial hydrogen ion concentration.
AID
JSR 4701
/
m4999$$124
02-20-96 13:48:18
FIG. 5. Arterial base excess in both groups of pigs.
srga
AP: Surg Res
314
JOURNAL OF SURGICAL RESEARCH: VOL. 61, NO. 2, MARCH 1996
FIG. 8. Clotting times in both groups of pigs.
FIG. 6. Fibrinogen levels in both groups of pigs.
filters the amniotic fluid before injecting it, the fluid is harmless [33]. It was further postulated that the mechanism of multiple organ failure in septic shock was the obstruction of visceral capillaries by microclots of DIC [3]. If this were true, the dissolution of the microclots would restore circulation to the organs. In this experiment, the administration of plasminogen activator resulted in a significantly decreased mortality and improvement in a number of other parameters, especially pH. It is difficult to prove that DIC actually occurs. Fibrin clots are sometimes found in autopsy material, but often are not. Many of the control pigs showed microscopic fibrin thrombi, as has been seen in other experiments [26]. It is theorized that the body’s fibrinolytic mechanism dissolves many of the microclots immediately after death. Perhaps the best indicator of DIC is the elevation of fibrin split products which occurred in all pigs. A theoretical danger of administering a plasminogen activator in the presence of shock is the development of a clotting defect and bleeding. This did not occur. In fact, the only clotting time increase was in a control pig. Clotting parameters otherwise
remained normal. In fact, fibrinogen was actually higher in the treated animals. Fibrinogen has been shown to increase very rapidly in response to any stress [23]. The fibrinogen level in the blood is determined by the balance between fibrinogen production and conversion to fibrin. In spite of using up of the fibrinogen by the DIC process, it has been found that in most cases of clinical DIC, fibrinogen is high [24, 25]. Many drugs have been tried to treat septic shock. Of course antibiotics, respiratory support, metabolic support by fluids and electrolytes, and nutritional support are all essential and universally used. However, mortality from septic shock, particularly due to multiple organ failure, especially adult respiratory distress syndrome, remains very high (50 to 90%) [27–34]. Corticosteroids have been extensively used with generally poor results [34, 35]. Vasoactive drugs, including vasoconstrictors, vasodilators, and mixed types have been helpful [36]. Endogenous opiates have been used with mixed results [37]. Many centers have been using monoclonal antibodies with great success in the laboratory [38–40] and also in some clinical studies [39, 40]. However, in an FDA-approved multicenter study, many unexpected deaths occurred which resulted in the removal from the market of interleukin-1 receptor antagonist [39]. A recent study of monoclonal antibody treatment showed an actual increase in mortality in canine gram-negative septic shock [41]. None of the above drug treatments have been shown to alter the high mortality in septic shock. Death from septic shock in the United States increased from 35,000 in 1970 to
FIG. 7. WBC in both groups of pigs.
AID
JSR 4701
/
m4999$$124
02-20-96 13:48:18
FIG. 9. Urinary output.
srga
AP: Surg Res
HARDAWAY, WILLIAMS, AND SUN: TREATMENT OF Escherichia coli
94,000 in 1980 to 198,000 in 1990 [42]. The death rate from septic shock per 100,000 population increased from 1.7 in 1970 to 4.2 in 1980 to 7.9 in 1990, according to the Statistical Abstract of the United States, 1992 [42]. For years investigators have focused on the pivotal role of endotoxin. It is composed of an O-polysaccharide chain, a core sugar, and a lipophilic fatty acid. The lipid a portion of the lipopolysaccharide includes expression of cytokine genes through stimulation of receptors on the surface of target cells and activation of the nuclear transcription factor NF-KB. This gene activation brings on the inflammatory response and its attendant deliterius effects, i.e., hypotension and organ disfunction. Because it causes a shock-like state when injected into experimental animals, endotoxin has been implicated as a causative agent in septic shock. However, as the understanding of sepsis is now more detailed, it is clear that endotoxin is only one component of sepsis. This may explain why trials in which antiendotoxin antibodies were injected into patients have failed to show consistent benefit.
9. 10.
11.
12.
13. 14. 15.
16. 17. 18.
CONCLUSION
Killed Escherichia coli organisms given intravenously cause death, with a low arterial and venous pH, high cardiac output, increased core temperature, elevated white cell count, and other metabolic changes. Death and the above changes in parameters can be significantly prevented by a plasminogen activator. Administration of a plasminogen activator in the presence of Escherichia coli shock causes no changes in the clotting mechanisms; in fact, it seems to prevent these changes. These results are consistent with the theory that DIC is produced by Escherichia coli organisms, that DIC causes the production of microclots which occlude the microcirculation of the viscera, and that a plasminogen activator will cause dissolution of these microclots with improvement in the circulation to the viscera.
19.
20. 21.
22.
23. 24.
25. 26.
REFERENCES
27.
1. Landois Transfusion des Blutes, Lypzig, 1875. 2. Schneider, C. L. Fibrin embolism with defibrillation as one of the end results during placenta abruptio. SGO 92: 1951. 3. Hardaway, R. M., Mckay, D. G., and Williams, J. H. Lower nephron nephrosis. Am. J. Surg. 87: 41–49, 1954. 4. Hardaway, R. M., and McKay, D. G. Production of lower nephron nephrosis in dogs by means of an episode of intravascular clotting. Surg. Form. IX: 134–137, 1959. 5. Hardaway, R. M., and McKay, D. G. Intravascular thrombi and the intestinal factor of irreversible shock. Ann. Surg. 150: 261– 265, 1959. 6. Hardaway, R. M., and McKay, D. G. Pathologic study of intravascular coagulation following incompatible blood transfusion in dogs. Am. J. Surg. 91: 24–40, 1956. 7. Hardaway, R. M., and McKay, D. G. Production of acute hemorrhagic pancreatitis in dogs by means of an episode of DIC. Surgery 45: 557–561, 1959. 8. Hardaway, R. M. Pulmonary microthrombi. In A. A. Sasahara
AID
JSR 4701
/
m4999$$124
02-20-96 13:48:18
28.
29.
30. 31.
32.
33.
34.
315
(Ed.), Pulmonary Embolic Disease, pp. 65–72. New York: Grune-Stratton, 1965. Hardaway, R. M. Disseminated intravascular coagulation syndromes. Arch. Surg. 83: 842–850, 1961. Hardaway, R. M. Syndromes of Disseminated Intravascular Coagulation with Special Reference to Shock and Hemorrhage. Springfield, IL: Charles C Thomas Pub., 1966. Hardaway, R. M., James, J. M., and Anderson, R. W., et al. Intensive study and treatment of shock in man. JAMA 199: 779–790, 1967. Hardaway, R. M., Husni, E. A., and Geever, E. F., et al. Endotoxin shock, a manifestation of intravascular coagulation. Ann. Surg. 154: 791–802, 1961. Hardaway, R. M. Mechanism of traumatic shock. SGO 151: 65– 69, 1980. Hardaway, R. M., and Williams, C. H. A new treatment of traumatic shock and ARDS. Resuscitation 19: 61–75, 1990. Hardaway, R. M., Hasni, E. A., Eseever, E. F., et al. Endotoxin shock, a manifestation of intravascular coagulation. Ann. Surg. 154: 791–802, 1961. Hardaway, R. M., Clovitz, H. O., Brewster, W. R., et al. Clotting time of heparanized blood. Arch. Surg. 89: 701–705, 1964. Hardaway, R. M., and Johnson, D. G. Influence of fibrinolysis on shock. JAMA 183: 397–599, 1963. Hardaway, R. M., and Burns, V. W. Mechanism of action of fibrinolysis in the prevention of irreversible hemorrhagic shock. Ann. Surg. 157: 305–309, 1963. Wall, T. C., et al. Results of high dose intravenous urokinase for acute myocardial infarction. Am. J. Cardiol. 65(3): 124–131, 1990. Nadrowski, L. F. Deep venous thrombosis: Recent advances in pathogenesis and treatment. Surg. Ann. 23(1): 1991. Hardaway, R. M., Williams, C. H., and Marvasti, M., et al. Prevention of adult respiratory distress syndrome with plasminogen activator in pigs. Crit. Care Med. 18: 1413–1418, 1990. Hardaway, R. M., and Williams, C. H. Prevention of multiple organ failure with plasminogen activator. Current Ther. Res. 49: 711–722, 1991. Hardaway, R. M., Johnson, O. G., et al. Influence of stress on fibrinogen. J. Trauma 4: 673–676, 1964. Hardaway, R. M., James, P. M., Anderson, R. W., et al. Intensive study and treatment of shock in man. JAMA 199: 779–790, 1967. Hardaway, R. M. The significance of coagulation and thrombotic changes after injury. J. Trauma 354–357, 1969. Hardaway, R. M. Syndromes of Disseminated Intravascular Coagulation. Springfield, IL: Charles C Thomas, Pub. 1966. Hardaway, R. M. Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. Hardaway, R. M. Treatment of adult respiratory distress syndrome with plasminogen activator. Hematol. Rev. Commun., in press. Hardaway, R. M., and Williams, C. H. Prevention of multiple organ failure with plasminogen activator. Curr. Ther. Res. 49: 711–722, 1991. Hardaway, R. M., and Johnson, D. Clotting mechanism in endotoxic block. Arch. Int. Med. 112: 775–782, 1963. Hinshaw, L. B., and Beller, B. K. Septic shock, physiological aspects. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publisher, 1988. Archer, L. T. Septic shock, pathological aspects. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. Schumer, W. Septic shock—Clinical aspects. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. Schumer, W. Treatment of septic shock with steroids. In R. M.
srga
AP: Surg Res
316
JOURNAL OF SURGICAL RESEARCH: VOL. 61, NO. 2, MARCH 1996 Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988.
35. Shatney, C. H. The role of corticosteroids in the treatment of shock. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. 36. MacConnell, K. L. Vasoactive drugs in the treatment of shock. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. 37. Bowen, J. C., and Rees, M. Endogenous opiates in shock. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. 38. Gaffin, S. L. Antibody therapy in shock. In R. M. Hardaway (Ed.), Shock—The Reversible Stage of Dying. Chicago, IL: Year Book Publishers, 1988. 39. Dudley, M. N. Interleukin-1 receptor antagonists significantly
AID
JSR 4701
/
m4999$$124
02-20-96 13:48:18
40. 41.
42. 43. 44. 45. 46.
reduce death from sepsis, University of Rhode Island MultiCenter Study, Gen. Surg. News Feb.: 28–29, 1993. Editor’s Note, Gen. Surg. News Feb.: 28, 1993. Quezado, Z. M. N., Natnason, C., Assinf, D. W., et al. A controlled trial of HA-IA in a canine model of gram-negative septic shock. JAMA 269: 2221–2227, 1993. Statistical Abstract of the United States, 1992. Hardaway, R. M. ‘‘Trauma, sepsis and disseminated intravascular coagulation.’’ J. Intensive Care Med., in press. Hardaway, R. M. Broken cell walls. The ‘‘toxin’’ of traumatic and septic shock. Critical Care Int., in press. Tuomanen, E. Breaching the blood-brain barrier. Sci. Am. Feb.: 80–84, 1993. Hardaway, R. M., and McKay, D. G. Disseminated intravascular coagulation: A cause of shock. Ann. Surg. 149: 462–470, 1959.
srga
AP: Surg Res