Search

All phenomena seen under ovine endotoxemia or bacteremia are typically observed in septic humans as well. The lethality (approximately 20%) in both sepsis models underlines the severity of the experimental sepsis in these models. As mentioned above, both models are ideal to objectify the effects of new therapeutic approaches for the treatment of sepsis, because they provide stable conditions. We tested the inhibition of nitric oxide synthase in both models: Nitric oxide is the main mediator of the vasodilation and the hyperdynamic circulation seen in sepsis. Since the restoration of the perfusion pressure is the major therapeutic goal to prevent further tissue damage (Chernow et al. 1990), the blockade of the nitric oxide synthase seems to be a logical approach for the treatment of hyperdynamic sepsis. Therefore, we tested the nitric oxide synthase inhibitor N(W)-nitro-L-arginine methyl ester (L-NAME) in the endotoxemic sheep model as well as in the bacteremic model. L-NAME reversed the hyperdynamic circulation of sepsis (Meyer et al.1992; Dehring et al.1993). The cardiac output was lowered back to baseline, and at the same time, the arterial pressure was elevated to baseline niveau, both resulting in a marked increase in systemic vascular resistance (figure 3). The pulmonary artery pressure showed only a slight increase, but due to the marked reduction in cardiac output the pulmonary vascular resistance increased significantly. The oxygen extraction was elevated to an extent, which prevented the oxygen consumption to fall, although the oxygen delivery dropped significantly because of the lowered cardiac output (figure 4). The intrapulmonary shunt was brought back to baseline (Meyer et al.1994a), allowing an improved pulmonary oxygen uptake. The renal function improved significantly after nitric oxide synthase inhibition in endotoxemia as well as in bacteremia (Hinder et al.1994; Lingnau et al.1994). Not only was the creatinine clearance elevated, but the urine output also increased, lowering the positive fluid balance. Another inhibitor of nitric oxide synthase N(W)-Mono-Methyl-L-Arginine (L-NMMA) was recently tested in these models as well. This drug is now already in clinical trials. The fact that the effects of these nitric oxide synthase inhibitors in septic humans are similar to the effects in the described experimental sepsis models proves the clinical relevance of the endotoxemic and the bacterimic sheep model.

The hemodynamic and lung lymph changes following near-drowning (ND) were studied in sheep. Experimental ND was by transtracheal aspiration of 10 ml/kg body weight of either seawater (SW) or freshwater (FW). Extravascular lung water and lung lymph protein flux were significantly increased, but cardiac index was depressed in both groups following ND. Systemic and pulmonary vascular resistances were markedly elevated with FW compared to only a slight rise with SW. Lung lymph oncotic pressure decreased with SW ND from baseline of 9.7 +/- 0.4 to 6.8 +/- 0.63 mm Hg (P < 0.05). In contrast, FW ND increased lung lymph oncotic pressure from 12.8 +/- 0.9 to 16.6 +/- 1.3 mm Hg (P < 0.05). These data suggest that the changes in lung lymph and hemodynamic response to SW and FW ND differ in sheep. The changes are immediate and profound with SW, but slower in onset and less severe with FW. FW ND is associated with hemolysis, which is absent in SW ND.

We have previously reported an increase in plasma levels of atrial natriuretic factor (ANF) in an ovine model of endotoxemia. The purpose of this study was to determine if this IR-ANF release was mediated by the increase of right atrial pressure (RAP) and right heart volumes concomitantly observed following endotoxin (LPS) administration. We studied right ventricular function, renal blood flow (RBF), urinary output (UO), urinary clearance of free water (CH20), urinary osmolality (UOSM), sodium excretion (UENA), and the plasma IR-ANF concentration (radioimmunoassay), following the administration of an E. coli LPS bolus (1 microgram/kg) with (group O, n = 8) and without (group E, n = 10) pretreatment with OKY-046, a selective thromboxane synthetase inhibitor. LPS induced early increases in RAP, right ventricular end-systolic (RVESV) and end-diastolic (RVEDV) volumes, heart rate (HR), and IR-ANF, and delayed increases in RBF, UO, and CH20. OKY-046 prevented the elevation of RAP, RVEDV, and RVESV; however, both groups showed virtually identical increases in IR-ANF (E: 20.03 +/- 3.8 to 192.33 +/- 35.47 pg/ml, O: 17.9 +/- 4.1 to 159.5 +/- 23 pg/ml) as well as an increase of HR, RBF, UO, and CH20. The increase in IR-ANF release noted following the administration of LPS in an ovine model does not appear to be related to the early elevations in right heart volumes or atrial distension.

For the past half-century, several high molecular weight compounds have been used for volume expansion during cardiopulmonary resuscitation. However, the effectiveness and side effects of these different expanders are varied. We have compared plasma, pentastarch, and a new product, pentafraction, for effective plasma volume expansion before and after tissue injury with endotoxin administration. In each group, eight range ewes instrumented with a Swan-Ganz, arterial, and venous catheters, and lung and flank lymphatic cannulas were compared. Each group received 15 ml/kg of either 6% pentafraction, 6% pentastarch, or plasma followed two hours later by 1.5 micrograms/kg/0.5 hr E. Coli endotoxin over 30 min. Data were collected for an additional 24 hr after endotoxin administration. Our results indicated a plasma volume expansion in all three groups. However, the prior administration of pentafraction significantly attenuated the increase in the lung lymph flow and early evaluation of systemic vascular resistance noted with endotoxin in comparison to the other two groups.

Eight awake sheep were monitored with ultrasonic crystals, positioned at the anterior and posterior left ventricular wall. A left-sided intraventricular pressure transducer and a right ventricular ejection fraction catheter were positioned in the right and left hearts, respectively. After administration of endotoxin (Escherichia coli, LPS 1.5 micrograms/kg in 30 min), the hemodynamic variables showed a triphasic course. Phase I, (0-1 hr post LPS) was characterized by an increased pulmonary artery pressure and a decreased right ventricular ejection fraction. The inability of the right ventricle to compensate for the increased preload resulted in a fall of the left ventricular preload, stroke volume, and cardiac output. Three hours after LPS administration a second drop of the cardiac output was noted (phase II). This occurred as a result of a fall in preload. Eight hours post LPS a hyperdynamic phase (phase III) was distinguished, with a high cardiac output and a low systemic vascular resistance. During this time there was evidence of probable reduced myocardial contractility.

We determined if the cardiopulmonary response to endotoxin (LPS) is mediated by bradykinin (BK). Sheep (n = 30) were prepared for chronic study with cardiopulmonary catheters, and chronic lung lymph fistulae. They were divided into the following study groups: A) BK infusion; B) LPS; C) LPS with angiotensin converting enzyme inhibitor (ACEI); D) ACEI alone; E) LPS with BK antagonist. Cardiac index and mean arterial pressure fell (6.9 +/- 0.4 to 5.3 +/- 0.3 L/min/m2 and 93 +/- 4 to 72 +/- 3 mmHg, respectively), and pulmonary lymph flow increased (10.5 +/- 3.0 to 17.8 +/- 4.3 ml/hr) during BK infusion. Addition of ACEI during BK infusion reduced the amount of BK necessary to induce a comparable response (125 to 2 micrograms/kg/hr). Administration of ACEI to LPS animals did not significantly alter their cardiopulmonary responses. BK antagonist did not change the response to LPS. Although the kallikrein-kinin pathway is believed to be activated during the septic response, our data do not lend support for the hypothesis that bradykinin is an important mediator of the cardiopulmonary changes.

To evaluate the effects of heparin on gram-negative septic shock, immature piglets were subjected to fecal Escherichia coli peritonitis. One group of animals received a continuous infusion of heparin 25 units/kg/hr, while the control animals were given an equivalent volume of lactated Ringer's solution. Heparin-treated animals (n = 5) had a mean survival time (+/- SEM) of 18.8 +/- 2.2 hr vs. 11.9 +/- 2.8 (P less than 0.05) in controls (n = 5). Animals receiving heparin tended to have improved hemodynamic profiles and less leukopenia than controls.

We have compared the effects of aspiration of 2.5 ml/kg of 0.1 N hydrochloric acid (HCl) on groups undergoing normal saline aspiration (2.5 ml/kg) and a sham control in unanesthetized sheep with chronic lung lymph fistula for up to 148 hr. Significant changes in pulmonary lymph flow and cardiopulmonary variables occurred in the airway acid group within the first 48 hr, as compared to saline and sham control. In the HCl group lymph flow increased from a baseline of 8.5 +/- 1.5 to 21.8 +/- 3.0 ml/hr (mean +/- SEM) at 2 hr. Pulmonary vasoconstriction occurred at this time with significant elevations of lymph thromboxane B2 from a baseline of 400 +/- 11 to 2,090 +/- 690 pg/ml. During the early acid aspiration phase products of lipid peroxidation were noted to increase from 0.29 +/- 0.03 to 0.81 +/- 0.17 absorbance units. Associated with elevated lymph flow was a significant drop in cardiac index in the airway-acid-aspirated group (6.59 +/- 0.52 to 5.53 +/- 0.36 L/min/m2) and a rise in bronchial blood flow from 9.8 +/- 1.7 to 63.7 +/- 8.9 ml/min. These data suggest that a moderate airway acid injury created by 2.5 ml/kg of HCl results in acute pulmonary damage with the release of several potential mediators which may be responsible for the observed effects. The long-term acid-injured animals survived for up to 1 week following the aspiration procedure with evidence of resolution of lung damage by wet-weight dry-weight evaluation of lung (normal lab control = 4.0 +/- 0.2, and 148 hr after acid = 4.5 +/- 0.3) and microscopic evidence of minimal persistent injury. These data document the pulmonary damage and repair resulting from airway acid instillation.

We compared the cardiopulmonary responses to a single bolus (1.5 microgram/kg) vs. continuous infusion of endotoxin (LPS) (24 ng/kg/hr) in unanesthetized sheep. A single bolus produced an initial marked increase in pulmonary arterial pressure and plasma thromboxane levels and an elevated flow rate of lung lymph low in protein. Concomitantly, the cardiac output dropped and systemic vascular resistance rose. In the animals that received a continuous infusion of LPS, only very small changes in these variables were noted during this early period. Later, lung lymph flow rate and protein flux were elevated in both groups with a greater response in the bolus group. At 6 hr after LPS, the systemic vascular resistance fell in both groups, but to a greater extent in the bolus group, whereas the cardiac output rose to the same extent. Plasma levels of neutrophils, lymphocytes, and plasma prekallikrein levels decreased in both groups; neutropenia was more pronounced in the bolus group. The most important difference between both endotoxemia models during this phase was the reduction of the stroke work index in the bolus model, which was not observed with the continuous infusion. The apparent myocardial depression, the early reduction in cardiac output, and eicosanoid mediated pulmonary hypertension are the major differences between the two responses.

The presence of myocardial depression is difficult to assess in most animal models of septic shock caused by hypovolemia. We studied endotoxemia in fluid-resuscitated conscious sheep. They underwent chronic instrumentation with pulmonary artery, arterial, and left atrial catheters. After 1 week of recovery, 1.5 micrograms/kg of endotoxin (LPS) was administered intravenously over a period of 30 min. One group of nine sheep stayed on baseline fluids (2 ml/kg/h of Ringer's lactate), while a second group (n = 6) was resuscitated with 7 ml/kg/h Ringer's lactate solution over the 24-h study period. Both groups were compared with controls. Eight hours after LPS, left ventricular stroke work index was depressed in both groups. In the group resuscitated with 7 ml/kg Ringer's lactate this depression associated with a normal left atrial pressure and a constant peripheral resistance indicated a shift in the Starling work curve downward and to the right evidencing myocardial depression.

The administration of endotoxin produces an early increase in hydrostatic pressure and pulmonary lymph flow, which is associated with elevated levels of thromboxane A2 in the lymph and can be blocked by the cyclooxygenase inhibitor ibuprofen. Two hours after the administration of endotoxin a secondary response is seen. The pulmonary lymph flow is elevated in association with a change in microvascular permeability to protein, this phase of the response can be reduced by the administration of the proteolytic enzyme inhibitor gabexate mesilate. In the present study we administered both compounds to eight sheep prepared for chronic cardiopulmonary and lung lymph studies to reduce both phases with the drug combination. The phase I response was reduced but the changes in lung lymph flow, associated with the phase II response, were unaffected by treatment. It is concluded that the effect of gabexate mesilate on the secondary response to endotoxin must in some way be related to the release of prostanoids.

Inhalation injury was produced in sheep that were chronically prepared for study. The injury was induced by insufflating them with smoke from burning cotton cloth. One group of animals was treated with the oxygen-free radical scavenger dimethylsulfoxide (DMSO) and heparin. Another group received heparin treatment alone, and a third was untreated. The drugs were nebulized into the tracheostomy at 4-hr intervals beginning 1 hr after injury. Following the inhalation injury, lung lymph flow and extravascular lung water measured by thermal-dye dilution technique were both increased. These elevations were associated with minor increases in pulmonary artery pressure, and, thus, since the lymph to plasma protein ratio was unchanged, this increased extravascular fluid formation was probably the result of an elevated microvascular permeability. These changes were associated with a reduction in alpha 2 macroglobulin antiprotease activity. The treated groups showed much smaller responses to the inhalation insult. This was especially true in the animals that received the DMSO. These findings support the concept that oxygen free radicals are responsible for the pulmonary edema associated with inhalation injury.

Prostaglandins released during inflammatory reactions cause increases in microvascular hydrostatic pressure, a primary cause of edema. Ibuprofen, a nonsteroidal, anti-inflammatory agent that reduces prostaglandin synthesis via inhibition of cyclooxygenase, was used to investigate the possible role of prostaglandins in the cardiopulmonary responses during sepsis. Sheep, surgically prepared for cardiopulmonary studies and collection of lung lymph, were given 0.75 micrograms/kg per 30 min of E. coli endotoxin iv. Ibuprofen (14 mg/kg) was given 15 min before and 1 h 45 min after the administration of endotoxin. We had previously noted a triphasic character to the hypovolemia encountered in endotoxin sepsis. The initial phase occurs during the first hour after endotoxin administration; it is characterized by decreases in PaO2, neutrophil count, and lymph-to-plasma (L/P) protein concentration ratios and by increases in mean arterial pressure, body temperature, hematocrit, lymph flow, and total plasma protein concentration. In the second phase these variables return toward their baseline values. In Phase 3 the same changes are observed an in Phase 1 except for a decrease in total plasma protein concentration and an increase in L/P ratios. Ibuprofen administration results in a statistically significant reduction in magnitude of Phase 1 changes, without notable effect on Phase 2 or Phase 3 values. These observations support the hypothesis that prostaglandins released during inflammatory reactions contribute to the extravascular fluid movement. Ibuprofen appears to lessen the severity of microvascular hydrostatic pressure-induced edema and the hypovolemia that occurs in the early stages of endotoxin.

Inhalation injury was produced in sheep which were chronically prepared for study. The injury was induced by insufflating them with smoke from burning cotton cloth. One group of animals was treated with the cyclooxygenase inhibitor ibuprofen and another group was untreated. Eight hr following the administration of smoke, there was an elevation of lung lymph flow in both groups. These changes were not as severe in the animals which were treated with ibuprofen. The pulmonary changes which occur following smoke inhalation injury are associated with elevations of the metabolites of arachidonic acid, especially those generated by the cyclooxygenase pathway. These metabolites in some way contribute to the pathophysiological changes induced by the inhalation of smoke, since they are reduced by the administration of a cyclooxygenase inhibitor.

Hyperdynamic sepsis (increased cardiac output and reduced peripheral vascular resistance) was created in sheep with chronic lung lymph fistulae (n = 8) by giving them a 30-min infusion of 1.5 micrograms/kg of endotoxin (LPS) iv. Four hours after LPS the cardiac output (CO) was reduced (6.56 +/- 0.43 to 4.96 +/- 0.33 liters/min) and lymph flow was increased (5.4 +/- 1.0 to 18.6 +/- 3.1 ml/h). Nine hours after LPS the CO output was increased (8.42 +/- 0.60 liters/min). Early cardiopulmonary changes were associated with a fall in neutrophils (PMNs) (2,667 +/- 748 to 450 +/- 90 cells/microliter) and an elevation of their chemiluminescence (CL), an indication of increased O2 free-radical formation in the blood (1,250 +/- 160 to 3,340 +/- 744 units/1,000 leukocytes). The granulocytic enzyme, aryl sulfatase, was increased in the lymph (0.19 +/- 0.03 to 0.37 +/- 0.05 microgram/h/mg protein) indicating degranulation (activation) of PMNs. When CO was increased (9 h after LPS), blood CL rose even higher (5,330 +/- 173 units/1,000 leukocytes) and CL in the lung lymph decreased (1,160 +/- 220 units/1,000 leukocytes). At this time, lymphatic aryl sulfatase had returned to baseline levels (0.25 +/- 0.02 microgram/h/mg protein). These data suggest that pulmonary microcirculatory injury produced by LPS may be the result of margination of PMNs in the lung and their release of permeability-inducing mediators. Later, as the CO increases, the PMNs or their lesion-producing mediators may be washed from the lung and the lung injury thus may be made less severe.

Inhalation injury was induced in chronically instrumented sheep (n = 9) by insufflating them with smoke from burning cotton cloth. Sham animals (n = 9) were insufflated with air. There were no temporal changes in any measured parameter of the sham animals. Smoke induced a depression in PaO2. There was a threefold elevation in protein-rich pulmonary lymph which was sustained for over 48 hours. The lymph-to-plasma oncotic pressure ratio was increased. The cardiac index, left atrial pressure, and pulmonary arterial pressure remained unchanged in both groups. After smoke inhalation, the interstitial levels of neutrophils increased while interstitial antiprotease activity was depressed. The lung lymph concentration of 6-keto prostaglandin F1a, the major metabolite of prostacyclin, was increased. These data suggest that the pulmonary injury following smoke inhalation is the result of an increase in lung microvascular permeability to protein with resultant pulmonary edema. The mechanisms responsible for these changes appear to be related to direct injury to the tracheobronchial tree by cytotoxic agents in the smoke; polymorphonuclear leukocytes; and, possibly, eicosanoids.

The cardiopulmonary effects of intravenously administered Escherichia coli endotoxin were studied in unanesthetized sheep. One group of animals was depleted of circulating T-lymphocytes while a non-depleted group served as control. T-lymphocyte depletion was accomplished by chronic thoracic duct drainage of lymph, removal of lymphocytes by continuous flow centrifugation and return of the cell free lymph intravenously. The T-lymphocyte depleted sheep demonstrated markedly obtunded increases in pulmonary arterial pressure and pulmonary vascular resistance following endotoxin when compared to the effects of the lipopolysaccharide in control animals. additionally, the lymphocyte depleted group showed a significant augmentation of myocardial contractility which occurred at the same time as marked systemic hypotension. This period of extreme hypotension following endotoxin is presumed to be accompanied by a reflex increase in the activity of the sympathetic nervous system. The control sheep, although equally hypotensive at this time, did not demonstrate a significant increase in myocardial contractility from the preendotoxin value. The results of these experiments indicate that T-lymphocytes may mediate some of the pathophysiological effects of bacterial endotoxin on the cardiovascular system.

Studies to test the effects of partial ganglionic blockade on renal and cardiovascular function were carried out in 16 mongrel dogs that under chloralose anesthesia had been subjected to full thickness flame burns to approximately 25% of their body surface. All animals received intravenous fluid replacement according to the same resuscitation formula we use in burned children. Half of the animals received 0.3 mg/kg of chlorisondamine hydrochloride 40 minutes after the burn; the remaining 8 dogs received only the vehicle. Among the variables monitored before burning and before and after blockade were glomerular filtration rate, renal plasma flow, water and osmolar clearance, sodium and potassium excretion, cardiac output, mean arterial, right atrial, and left ventricular end diastolic pressure, peripheral resistance, peak dP/dt/P, blood pH, and blood gases. Analysis of the data has revealed that pharmacologic blockade of the sympathetic system during the immediate postburn period results in a marked improvement in cardiac output and moderate improvement in kidney function.

Peritonitis-induced septic shock in the neonate is associated with a high mortality. Because there exists no clinically relevant model to study resuscitation of these patients, a model using the neonatal pig was developed. After arterial and central venous cannulation, and placement of a left pulmonary-artery thermodilution catheter, 12 anesthetized neonatal pigs were "resuscitated" with fluids (5% albumin in lactated Ringer's solution at 15 ml/kg/hr), antibiotics, and correction of acidemia. The pigs were divided into two groups: a control group (n = 5), which was not subjected to peritonitis and which was killed after 6 hours of monitoring, and a septic group (n = 7), which was inoculated with an intraperitoneal injection of sterile pig feces and Escherichia coli and was monitored until death (mean survival time (S.D.) 546(159) minutes). Serial measurements of hemodynamic and laboratory data were obtained. While pigs in the control group showed no significant changes in these data as measured against time, the pigs in the experimental group showed an early transient rise in cardiac index which was significant (p less than .05) and which was followed by a steady decline in cardiac index until death. These changes in cardiac index were accompanied by a continuous decline in mean arterial pressure, central venous pressure, pulmonary artery pressure, and systemic vascular resistance index, while pulmonary vascular resistance index showed a gradual continuous rise. The observed changes in hemodynamic and laboratory data mimic those anticipated in the human neonate with peritonitis-induced septic shock. This model proves reliable and reproducible, and shows promise as a tool to study the resuscitation of neonates with septic shock.

The mortality rate of inhalation injury was lowered in six chronically instrumented sheep through positive pressure ventilation and positive end-expiratory pressure (PEEP). Six range ewes were prepared for study by implanting catheters to measure lung lymph flow and cardiopulmonary variables. After surgery, these animals were studied in the unanesthetized state and then subjected to an inhalation by insufflating them with smoke from burning cotton. Following the smoking procedure, the animals were studied for 72 hr. All had marked falls in arterial oxygen tension and they developed dyspnea within 24 hr. The inhalation injury produced a marked change in lung lymph flow concomitant with an elevation in the lymph to plasma (L/P) oncotic pressure ratio. This is characteristic of a change in microvascular permeability to protein. A tracheostomy was performed at 72 hr and the animals were connected to positive pressure ventilators with PEEP. All six animals survived. It was concluded that the sheep lung lymph preparation is a very suitable model for the study of inhalation injury and positive pressure ventilation.

Sheep were infused with live bacteria to determine if the bacteria are phagocytized in the pulmonary circulation and to study the associated cardiopulmonary changes. Unanesthetized animals (n = 9) with chronic hemodynamic and pulmonary lymph catheters received a 1 hr central venous infusion of live Pseudomonas aeruginosa (5 x 10(7) Ps./minute) and were compared to a sham group (n = 7). The pulmonary arterial levels of bacteria were five to 100 times higher than the aortic levels. Pulmonary intravascular clearance rates were 79-91%. Electron microscopy of the lungs 24 hr after the bacterial infusion showed that pulmonary intravascular macrophages and neutrophils phagocytosed the bacteria. Severe initial and mild persistent pulmonary hypertension occurred. The pulmonary lymph flow was elevated, initially from hydrostatic pressure and later from increased permeability. A hyperdynamic circulation occurred from 6 to 18 hr, with elevated cardiac index and lowered systemic vascular resistance and mean arterial pressure, mimicking cardiopulmonary changes seen in clinical sepsis. The removal of bacteria in the lungs may contribute to the injury in sheep.

Sheep which have been previously prepared for cardiopulmonary studies and collection of lung lymph were given 0.75 micrograms/kg of E coli endotoxin iv. This induced sepsis produced a triphasic response. Phase 1 occurs during the first hour after endotoxin administration and is characterized by a decreased cardiac output, lymph to plasma protein ratio, neutrophil count, and an increased hematocrit, total plasma protein concentration, lymph flow, and pulmonary artery pressure. During phase 2 these variables tend to return toward their baseline values. Phase 3 begins 2.5 hr after the administration of endotoxin and shows many of the same changes that were observed in phase 1. Also during the late phase, the plasma protein concentration and lymphocyte count were reduced and the lymph to plasa protein ratio was increased. It is concluded that (1) An early fall in cardiac output occurs as a consequence of hypovolemia. The decreased volume is the result of fluid movement from the vascular compartment to the interstitial space consequent to a microvascular pressure increase. Since the changes occur coincidentally with a drop in neutrophil count, these cells may bear some causal relationship to the response. (2) The late fall in cardiac output in phase 3, also the result of a diminished vascular volume, occurs secondarily to extravasular fluid movement as a consequence of both an elevated microvascular pressure and an increased permeability to protein. The latter may be causally related to a fall in lymphocytes.

Peptidoglycans (PG) and lipopolysaccharides (LPS) are cell wall constituents of bacteria. The relative contents vary significantly between gram-negative and gram-positive strains. Since both strains take part in clinical septicemia, we compared hemodynamic and permeability properties in a sheep model with chronic instrumentation and lung lymph drainage. The cardiovascular reaction patterns of PG and LPS were comparable at very different dose levels, with LPS being 10,000-fold more potent. Continuous infusion produced an increase in lung permeability (only with LPS) and cardiac output, as well as fever, to mimic the hyperdynamic situation in early septicemia.