The Benefits of Corticosteroids

Transcription

1 The Benefits of Corticosteroids in Endotoxic Shock Richard Prager, M.D., Marvin M. Kirsh, M.D., Ernest Dunn, M.D., Ronald Nishiyama, M.D., John Straker, B.S., Robert Lee, B.S., and Herbert Sloan, M.D. ABSTRACT The experiments reported here were undertaken to study the effects of pharmacological doses of corticosteroids administered alone or in conjunction with prolonged ( 12-hour) assisted circulation in 22 dogs subjected to LD,,, Escherichia coli endotoxin. The most striking findings were lengthened survival time, higher cardiac output, decreased fluid requirement, and minimal evidence of pulmonary congestion or injury in the animals treated with steroids only. Unexplained mesenteric infarction prematurely terminated the experiments in animals undergoing assisted circulation. The benefits of corticosteroids in experimentally induced endotoxic shock are clearly demonstrated in these experiments. Further studies are needed to clarify the supportive role of assisted circulation in endotoxic shock and to determine any possible advantage of hypothermia over normothermia during its course. D espite a better understanding of the nature of the physiological and cellular abnormalities associated with septic shock as well as improvements in therapeutic measures, shock as a complication of sepsis is still a major cause of in-hospital deaths in the United States. Approximately 70,000 patients die yearly as a result of this grave complication [9]. Results from both experimental and clinical studies have implicated myocardial depression as an important etiological factor in the pathogenesis of septic shock [6-8, 131. Sepsis also imposes an increased metabolic requirement that is normally met in part by an increase in cardiac output. The diseased heart, however, may not possess sufficient functional reserve to meet this additional requirement. As a result there is greater compromise of myocardial function, leading to a further decrease in cardiac output. It has also been shown by several investigators that patients in septic shock have a diminished oxygen uptake, with resultant tissue hypoxia leading to even further impairment of myocardial function [I, 31. Theoretically, prolonged circulatory assistance with a pulsatile pump in From the Departments of Surgery and Pathology, The University of Michigan Medical Center, Ann Arbor, Mich. The authors wish to thank Jeanne Tashian for her editorial assistance in the preparation of this manuscript. Accepted for publication Oct. 21, Address reprint requests to Dr. Kirsh, C-7175 University Hospital, Ann Arbor, Mich THE ANNALS OF THORACIC SURGERY

2 Corticosteroids in Endotoxic Shock combination with a membrane oxygenator can help patients in septic shock by improving oxygen delivery to the cells as well as decreasing the workload of the heart. Previous reports from our laboratory [4] have demonstrated the supportive role of pulsatile assisted circulation with a membrane oxygenator in experimentally induced endotoxic shock. Twelve animals given LDloo endotoxin developed acute renal failure that led to their death soon after cessation of assisted circulation. Hypothermia, by creating a more favorable balance between oxygen supply and demand, and pharmacological doses of corticosteroids, whose exact mechanism of action is unknown, have also been reported to be of benefit in patients with septic shock. The purpose of these experiments was to study the effects of pharmacological doses of corticosteroids alone or in combination with prolonged (12-hour) assisted circulation using a pulsatile assist pump and membrane oxygenator in dogs subjected to LD50-60 Escherich ia coli endotoxin. Methods Twenty-two adult mongrel dogs weighing between 25 and 35 kg were anesthetized intravenously with sodium pentobarbital (30 mg per kilogram of body weight). A cuffed endotracheal tube was inserted and ventilation accomplished with a positive-pressure respirator using 98y0 oxygen and 2y0 carbon dioxide. When necessary, curare (1 ml) was administered intravenously to control movement. Ventilation was monitored by periodic measurements of Po, and Pco2 and adjusted as required. Whenever the animal exhibited a Po, less than 60 mm Hg, positive end-expiratory pressure was applied. A polyethylene catheter with a thermistor attached was placed in the aortic arch through the left carotid artery, and a Swan-Ganz balloon catheter was introduced through the right femoral vein and guided into the main pulmonary artery. Right atrial, systemic arterial, and pulmonary arterial pressures were continuously monitored with Statham pressure transducers. Cardiac output was determined by the thermodilution technique. Urine output was determined hourly. Beef lung heparin" (1 mg/kg) was administered intravenously. Hemodynamic studies were conducted before the intravenous administration of a single LDoo dose of E. coli endotoxinj- (0.5 mg/kg) and at 60-minute intervals afterward. In addition, pulmonary static compliance utilizing a bellows pressure apparatus and pulmonary shunting using Berggens equations were determined at baseline and at 3-hour intervals thereafter. The animals were then randomly divided into four groups. +Upjohn Corp., Kalamazoo, Mich. tdifco Laboratories, Detroit, Mich. VOL. 19, NO. 2, FEBRUARY,

3 PRAGER ET AL. GROUP 1: CONTROL Treatment was begun in these 6 animals immediately after endotoxin injection by maintaining circulating blood volume and acid-base balance. Ringer s lactate solution was used if the hematocrit was greater than 35y0 and unmatched whole blood if it was less. One or the other was infused until the arterial pressure reached 90 mm Hg. Sodium bicarbonate (0.3 mg/kg/meq HCOa deficit /liter) was given intravenously by the hour to correct metabolic acidosis. Furosemide in graduated doses from 0.5 up to 7.0 mg/kg was started if the animal had 2 consecutive hours of anuria. Serum potassium was maintained between 3.5 and 5.5 meq/liter by the intermittent intravenous administration of potassium chloride. Kanamycin was given through a nasogastric tube at hour 4 and hour 8. Heparin (0.5 mg/kg was administered hourly to maintain the thrombin clotting time between 30 and 40 seconds. This regimen was maintained for 12 hours. At its conclusion the animals were extubated, observed for several hours, and then returned to their cages. At the end of 48 hours the survivors were killed and subjected to postmortem examination. Animals not surviving the 48-hour period underwent postmortem examination immediately after their death. Gross and microscopical examinations of the heart, lungs, kidneys, and gastrointestinal tract were performed. GROUP 2: STEROID ONLY These 6 animals were treated identically to those in Group 1 except that they received an intravenous bolus of methylprednisolone sodium succinate (30 mg/kg)* at 1 and 7 hours following the injection of E. coli endotoxin. GROUP 3: STEROIDS AND ASSISTED CIRCULATION-NORMOTHERMIA In these 6 animals the superior and inferior venae cavae were cannulated with Bardic catheters introduced through the external jugular vein and femoral vein, respectively, and the common iliac artery was cannulated with a No. 18 Bardic catheter advanced into the descending thoracic aorta. Venous drainage and arterial perfusion were achieved with these cannulas. Treatment with the pulsatile assist pump and membrane oxygenator was begun for 3 animals prior to endotoxin injection and for 3 animals 60 minutes after endotoxin injection. A General Electric Dualung membrane oxygenator was used to provide oxygenation. A standard roller pump driven by an electronic circuit that allowed for synchronization during diastole was modified to deliver pulsatile flow. Normothermic partial cardiopulmonary bypass was maintained for 12 hours or until death at 60 to SOY0 of basal cardiac output. Cardiopulmonary bypass was discontinued 5 minutes each hour for hemodynamic and pulmonary function studies and blood gas determinations. At 1 and 7 hours following the injection of endotoxin, the dogs received an intravenous bolus of methylprednisolone sodium succinate (30 mg/kg). Sodium bicarbonate was given intravenously *Upjohn Corp., Kalamazoo, Mich. 144 THE ANNALS OF THORACIC SURGERY

4 Corticosteroids in Endotoxic Shock every hour to correct metabolic acidosis. Furosemide was administered exactly as in the Group 1 and Group 2 animals. Kanamycin (10 mg/kg) in saline suspension was given much as it had been in the Group 1 and Group 2 dogs. Serum potassium was maintained between 3.5 and 5.5 meq/liter by the intermittent intravenous administration of potassium chloride. Heparin (0.5 mg/kg) was given hourly to maintain the thrombin clotting time between 30 and 40 seconds. At the conclusion of cardiopulmonary bypass, the clotting mechanism was reversed and the animals were extubated and observed for several hours. All animals underwent postmortem examination immediately after death. Gross and microscopical examinations of the heart, lungs, kidneys, and gastrointestinal tract were performed. GROUP 4: ASSISTED CIRCULATION-HYPOTHERMIA These 4 animals were treated identically to those in Group 3 except that they underwent partial cardiopulmonary bypass at moderate hypothermia (28"-30 C) at flow rates between 60 and 80% of basal cardiac output. Results SURVIVAL The most striking finding was the lengthened survival time of the Group 2 animals. Four of the 6 lived for 48 hours following endotoxin injection, whereas only 2 of 6 Group 1 animals survived for this period of time. None of the 10 animals (Groups 3 and 4) that underwent partial cardiopulmonary bypass survived the observation period. During perfusion, 7 of these 10 animals developed extensive mesenteric infarction which caused their death, a phenomenon independent of the interval between endotoxin injection and the institution of cardiopulmonary bypass. In addition, there was no evidence of venous obstruction by the cannulas at postmortem examination. Two Group 4 animals survived the period of perfusion but died of unknown causes before the end of the 48-hour observation period. One Group 3 animal developed sustained bradycardia, hypotension, and cardiac arrest soon after cessation of assisted cardiopulmonary circulation. HEMODYNAMICS There was no statistically significant difference in any of the groups between the decrease in cardiac index, the systemic arterial pressure, or the changes in peripheral vascular resistance 2, 4, 8, and 12 hours following the injection of endotoxin (Figs. 1, 2). FLUID BALANCE The most consistent difference between Group 1 and Group 2 was the decreased amount of fluids required by the Group 2 animals to maintain an adequate systemic arterial pressure (> 90 mm Hg) during the period of observation. The Group 1 animals required an average of 110 ml/kg during the 1 2-hour period, whereas the Group 2 steroid-treated animals required only 74 ml/kg during the same period (p < 0.01). Because of their decreased VOL. 19, NO. 2, FEBRUARY, 1!)75 145

5 PRAGER ET AL. a I I I 1 I 1 O b HOURS FIG. 1. Hemodynamic studies in non-steroid-treated (Group 1) and steroid-treated (Group 2) animals following administration of LDso-,,o E. coli endotoxin. (Dotted lines = non-steroid-treated [Group 13; solid lines = steroid-treated [Group 21.) fluid requirement, animals in Group 2 gained less weight than those in Group 1. Renal failure occurred in only 1 animal in each of these two groups. Massive fluid replacement was required in the 7 animals that developed mesenteric infarction while on cardiopulmonary bypass. PATHOLOGY Pulmonary injury or congestion was present in varying degrees in all the animals but was most severe in the control group (Group 1) and in the animals undergoing cardiopulmonary bypass (Groups 3 and 4). The lungs in these animals were heavy, edematous, and in some cases even 146 THE ANNALS OF THORACIC SURGERY

6 x 4.0 =\ Corticosteroids in Endotoxic Shock r I- T I- h I 1 I I 1 J 160 W u z 140 z v, mo gsi 120 A a\ fj? g: 80-, W a A P _ g - consolidated. Microscopical examination revealed widespread areas of interstitial and intraalveolar hemorrhage and edema, atelectasis, and periarterial hemorrhage (Fig. 3). In contrast to these findings was the minimal evidence of pulmonary congestion or injury in the Group 2 animals. Their lungs were slightly increased in weight and exhibited a minimal amount of edema on gross examination. Microscopically there were VOL. 19, NO. 2, FEBRUARY,

7 PRAGER ET AL. FIG. 3. Repesentative lung section from an animal in Group 1 demonstrates marked perivascular hemorrhage. (Hd.E, ~160; enlarged 25% for reproduction.) only focal areas of interstitial and intraalveolar hemorrhage and edema, atelectasis, and periarterial hemorrhage. Examination revealed no difference among the four groups with respect to the heart and kidneys. The hearts in all the animals exhibited scattered areas of epicardial and endocardia1 hemorrhage. These changes were most marked in those animals which had been noticeably hypotensive for several hours prior to death. The kidneys on gross examination were edematous and showed scattered areas of subcortical hemorrhage. Microscopical examination demonstrated widespread or scattered areas of parenchymal hemorrhage. The most striking finding present in the animals undergoing cardiopulmonary bypass (Groups 3 and 4) was the mesenteric infarction that occurred in 7 of the 10 animals and involved both the small and large intestine. Gross and microscopical examination provided no ready explanation for its occurrence, nor was there evidence of venous or arterial obstruction by the perfusion cannula. Comment The effect of septic shock on the heart has been the subject of continuing controversy. Although the preponderance of data in the 148 THE ANNALS OF THORACIC SURGERY

8 Coiticosteroids in Endotoxic Shock literature supports the concept that some degree of myocardial deterioration may be present in septic shock, it is not certain whether the cardiac depression is primary or secondary in nature. Urschel and colleagues [17] and Greenfield and his associates [6] concluded that myocardial depression, when it occurs during the course of septic shock, results from pathogenic processes other than a direct effect of the endotoxin itself. Elkins and coworkers [5] demonstrated that coronary perfusion pressure is one of the major determinants of preservation of myocardial function in endotoxic shock. On the other hand, Hinshaw, Cann, and Solis and their colleagues-among others-have shown that primary myocardial depression plays a significant role in the pathophysiology of endotoxic shock. Solis and Downing [ 141 noted the early onset of progressive deterioration of left ventricular function following an injection of E. coli endotoxin. Hinshaw s group [7, 81, using an isolated working canine heart model, demonstrated the occurrence of heart failure 5 to 7 hours after endotoxin administration. Using intact animals, Cann and associates [2] also demonstrated a primary myocardial depressant effect of endotoxin. Patients in septic shock appear to have an abnormality in oxygen transport manifested by a reduced ability to extract oxygen. The exact mechanism for this abnormality is not clear [1, 31. Although these patients are usually febrile and in fact appear to be in a hypermetabolic state, their oxygen consumption is usually normal or decreased. The reduced oxygen consumption seems to be related either to toxic inhibition of cellular metabolism or to the increased pulmonary venoarterial admixture that occurs in these patients. In theory, assisted circulation may be of some advantage to patients in septic shock. First, the modified roller pump is capable of delivering pulsatile flow synchronized with diastole. This increases coronary blood flow by raising diastolic pressures, and it reduces left ventricular workload, thereby allowing preservation or recovery of myocardial function. In addition, during the period of assisted circulation there is improved tissue perfusion with less metabolic acidosis. Second, because venoarterial bypass with a membrane oxygenator is utilized, oxygen transport to the cells is improved and the oxygen demand of the body tissues can then be met. LD,,60 E. coli endotoxin was used in these experiments to avoid the problem of acute renal failure encountered in our previous studies, in which LDloo endotoxin was administered [4]. Also, in an attempt to avoid the many problems associated with 24-hour perfusion, the animals were placed on partial cardiopulmonary bypass for only 12 hours. None of the 6 animals undergoing normothermic pulsatile assisted circulation (Group 3) survived the 12-hour period of perfusion; 5 of the 6 developed extensive mesenteric infarction that caused their death. This occurred independent of the interval between the injection of E. coli endotoxin and the institution of cardiopulmonary bypass. Because of this unexplained occurrence, it was VOL. 19, NO. 2, FEBRUARY,

9 PKAGEK ET AL. impossible in these experiments to determine the beneficial role of normothermic pulsatile assisted circulation in experimentally induced endotoxic shock. The Group 4 animals underwent a period of assisted circulation at moderate hypothermia (28"-30 C). By reducing the metabolic rate while simultaneously maintaining an adequate cardiac output with assisted circulation, we hoped that a more favorable balance between oxygen supply and demand would develop, thereby preventing hypoxic injury at the cellular enzyme!evel. An additional advantage of hypothermia over normcsthermia is the diastolic augmentation that occurs in these animals because of their slower heart rate. Two of the animals in this group developed mesenteric infarction that prematurely terminated the experiment. The other 2 animals survived the period of circulatory assist only to die overnight from unknown causes. Although prolonged survival was not achieved, it was of interest that these 2 animals required less fluid during perfusion and had less evidence of intraalveolar hemorrhage and pulmonary edema on microscopical examination than the animals undergoing normothermic partial cardiopulmonary bypass. Further studies are needed to determine the role of hypothermia in conjunction with assisted circulation in the treatment of septic shock. The value of steroids in pharmacological doses for the treatment of septic shock remains a controversial issue. Rosenbaum [111, Schumer [12], and Lillehei [lo] and their colleagues have demonstrated that steroids are useful in septic shock. On the other hand, Thomas and Brockman [16] found that the administration of steroids did not alter survival or prolong life in experimental animals if given after the induction of endotoxic shock. Pretreatment with steroids before induction of shock increased survival but did not alter the usual hemodynamic response to endotoxin. A variety of hemodynamic and membrane-stabilizing effects have been postulated to explain the beneficial action in septic shock. These include: a positive inotropic effect with subsequent increase in cardiac output and blood pressure, increased venous return, decreased peripheral vascular resistance, potentiation of the circulatory effects of catecholamines, and stabilization of lysosomal membranes [2, Results of the present experiments clearly demonstrate the beneficial value of pharmacological doses of corticosteroids in experimentally induced endotoxic shock. Four of the 6 animals (Group 2) that received 30 mg/kg of methylprednisolone sodium succinate were long-term survivors, whereas only 2 of 6 identically treated animals (Group 1) that did not receive steroids survived for a prolonged period. There was no difference between the two groups with respect to systemic arterial pressure, cardiac index, percentage of intrapulmonary arteriovenous shunting, pulmonary arterial pressure, or systemic vascular resistance during the period of observation. These results are in aveement with those of Spath and associates [15], who also failed to find an acute positive inotropic effect of glucosteroids in shock, although 150 THE ANNALS OF THORACIC SURGERY

10 Corticosteroids in Endotoxic Shock increased survival was associated with their use. Similarly, our data failed to demonstrate that the steroids induced systemic vasodilation during the period of observation. Of interest was the fact that the animals in Group 2 required significantly less fluid to maintain an adequate systemic arterial pressure and urine output than those in Group 1 (p < 0.05). These data indicate that the beneficial effects of glucocorticosteroids in septic shock are not related to vasodilation or to positive inotropic effect but in fact result from their membrane-stabilizing effect. One can only speculate on whether or not stabilization of the lysosomal membrane inhibits the generation of a myocardial depressant factor, as hypothesized by Lefer [ 151. In further support of the membrane-stabilizing effect of corticosteroids is the fact that pulmonary cellular damage was prevented and the pulmonary microarchitecture in Group 2 animals subsequently preserved. Although pulmonary congestion was present in these animals, it was strikingly less than in Group 1 animals. The lungs in Group 2 animals showed less periarterial hemorrhage, pulmonary edema, and microatelectasis on microscopical examination than the lungs of Group 1 animals. These results are consistent with those previously reported by Wilson [18], who demonstrated that pharmacological dosage of methylprednisolone sodium succinate prevents injury to the type I1 pneumocyte and stabilizes the basement capillary membrane as well. References 1. Albrecht, M., and Clowes, G. H. A.. Jr. The increase of circulatory requirements in the presence of inflammation. Surgery 56: 158, Cann, M., Stevenson, T., Fiallos, E., and Thal, A. P. Depressed cardiac performance in sepsis. Surg Gynecol Obstet 134:759, Clowes, G. H. A., Jr. Oxygen Transport and Utilization in Fulminating Sepsis and Septic Shock. In S. G. Hershey, L. R. M. Del Guercio, and R. McGoon (Eds), Septic Shock in Man. Boston: Little, Brown, Dunn, J., Kirsh, M. M., Harness, J., Lee, R., Straker, J., and Sloan, H. The role of assisted circulation in the management of endotoxic shock. Ann Thorac Surg 17:574, Elkins, R. D., McCurdy, J. R., Brown, P. P., and Greenfield, L. J. Effects of coronary perfusion pressure on myocardial performance during endotoxin shock. Surg Gynecol Obstet 137:991, Greenfield, L. J., McCurdy, J. R., Hinshaw, L. B., and Elkins, R. C. Preservation of myocardial function during cross-circulation in terminal endotoxin shock. Surgery 72: 11 1, Hinshaw, L. B., Archer, L. T., Black, M. R., Greenfield, L. J., and Guenter, C. A. Prevention and reversal of myocardial failure in endotoxin shock. Surg Gynecol Obstet 136:1, Hinshaw, L. B., Archer, L. T., Greenfield, L. J., and Guenter, C. A. Effects of endotoxin on myocardial hemodynamics, performance, and metabolism. Am J Physiol 221:504, Hinshaw, L. B., Greenfield, L. J., Owen, S. E., Black, M. R., and Guenter, C. A. Precipitation of cardiac failure in endotoxin shock. Surg Gynecol Obstet 135:39, 1972.

11 PRAGER ET AL. 10. Lillehei, R. C., Dietzman, R. H., and Morac, S. Treatment of septic shock. Mod Treat 4:321, Rosenbaum, R. W., Hayes, M. F., and Matsumoto, T. Efficacy of steroids in the treatment of septic and cardiogenic shock. Surg Gynecol Obstet 136:914, Schumer, W., and Nyhus, L. M. The role of corticoids in the management of shock. Surg Clin North Am 49: 147, Siegel, J. H., and Fabian, M. Therapeutic advantages of an inotropic vasodilator in endotoxin shock. JAMA 200:696, Solis, R. T., and Downing, S. E. Effects of E. coli endotoxemia on ventricular performance. Am J Physiol211:307, Spath, J. A., Gorczynski, R. J., and Lefer, A. M. Possible mechanisms of the beneficial action of glucocorticoids in circulatory shock. Surg Gynecol Obstet 137:597, Thomas, C. S., Jr., and Brockman, S. K. The role of adrenal corticosteroid therapy in Escherichia coli endotoxin shock. Surg Gynecol 0 bstet 126:61, Urschel, C. W., Serur, J. R., Forrester, J. H., Amsterdam, E. H., Parmley, W. W., Dembetsky, W., and Sonnenshin, E. H. Myocardial Contractility during Hemorrhagic Shock, Endotoxemia and Ischemia. In B. K. Forschcr, R. D. Lillehei, and S. S. Stubbs (Eds), Shock in Low-Flow and High-Flow States. Amsterdam: Excerpta Medica, Pp Wilson, J. W. Treatment or prevention of pulmonary cellular damage wii:h pharmacologic doses of corticosteroids. Surg Gynecol Obstet 134:675, THE ANNALS OF THORACIC SURGERY