The role of leukotriene B4 in endotoxin-induced lung injury in unanesthetized sheep

In order to examine the role of LTB4, a potent neutrophil chemokinetic and chemotactic factor, in the lung injury induced by Escherichia coli endotoxin, we measured LTB4 in systemic arterial blood plasma and lung lymph in unanesthetized sheep with chronic lung lymph fistulas. E. coli endotoxin (1 microgram/kg) infusion produced a biphasic response. The early period (Phase 1) was a transient pulmonary hypertension. The late period (Phase 2) was a more prolonged period characterized by an increase flow of lung lymph with a high concentration of protein, suggesting increased pulmonary vascular permeability. Peripheral leukocyte counts rapidly decreased during Phase 1 and leukopenia persisted for approximately 5 h. The concentration of LTB4 in arterial plasma and lung lymph significantly increased during Phase 1, and then decreased with a rebound significant increase during Phase 2. That is, LTB4 in plasma and lung lymph showed a biphasic increase after endotoxin infusion. Our data suggest that the elevation of LTB4 is related to the pulmonary leukocyte sequestration in the lung and may contribute to the lung vascular injury induced by endotoxin in unanesthetized sheep.     

Reduction of total hemolytic complement activity with Naja haje cobra venom factor does not prevent endotoxin-induced lung injury in sheep

We studied the effects of reducing total hemolytic complement activity with Naja haje cobra venom factor on the lung injury caused by intravenously infused endotoxin in 5 unanesthetized sheep with lung lymph fistulas. In normal sheep, infusions of lipopolysaccharide W from Escherichia coli (1.0 micrograms/kg) intravenously over 30 min caused increases in protein-rich lung lymph flow as well as the appearance in plasma and lung lymph of complement (C5)-derived chemotactic activity for polymorphonuclear leukocytes. Reduction of total hemolytic complement activity by treatment with Naja haje cobra venom factor (12 to 17 U/kg intraperitoneally) did not prevent the lung injury caused by endotoxin and also did not prevent the appearance in plasma and lung lymph of chemotactic activity. We conclude that although complement appears to be activated following intravenously infused endotoxin in sheep, a completely intact complement system is not necessary for endotoxin-induced lung injury.     

Perilla ketone: a model of increased pulmonary microvascular permeability pulmonary edema in sheep

A model of increased microvascular permeability pulmonary edema was developed in chronically instrumented unanesthetized sheep using perilla ketone (PK). PK did not cause changes in pulmonary hemodynamics but did cause marked increases in the flow of protein-rich lung lymph. The changes in lung lymph flow were accompanied by radiographic evidence of both interstitial and alveolar pulmonary edema as well as hypoxemia. PK did not cause acute changes in lung mechanics. Dynamic compliance of the lungs and FRC decreased later, concomitant with the changes in lung lymph flow, radiographic evidence for pulmonary edema, and hypoxemia. Resistance to air flow across the lungs and specific conductance did not change significantly after PK infusion. The severity of the radiographic evidence for pulmonary edema observed after PK correlated with the severity of the concomitant hypoxemia and changes in dynamic compliance of the lungs. PK did not cause increases in the concentrations of cyclooxygenase products of arachidonic acid in lung lymph or plasma or changes in blood leukocyte counts. We conclude that PK causes increased lung microvascular permeability pulmonary edema without acute changes in pulmonary hemodynamics. This model permits study of the pathophysiologic aspects of increased lung microvascular permeability without the concomitant functional alterations that complicate most other experimental models of diffuse lung injury.     

Effect of hemorrhagic shock on endotoxin-induced pulmonary hypertension and increased vascular permeability in unanesthetized sheep

The lung is very susceptible to sepsis or endotoxin injury in the trauma patient. We studied the effect of an episode of hemorrhagic shock and resuscitation on the prostaglandin-induced pulmonary hypertension and leukocyte-induced increased permeability phase of endotoxin lung injury. Eight unanesthetized sheep with chronic lung lymph fistula were bled 50% of blood volume for 2 hr, then resuscitated. Thromboxane, TxA2, levels increased from 0.1 to 0.6 ng/ml during shock, while blood white cell count decreased. Both parameters returned to baseline while lung lymph flow increased twofold during resuscitation with lymph being protein-poor, indicating no increase in permeability. Lung water was not increased but some pulmonary leukostasis was evident histologically after resuscitation. We then studied the effect of this process on all immediate endotoxin insult. Seven unanesthetized sheep were given 0.7 microgram/kg E. coli endotoxin alone, and again after shock and resuscitation, in paired studies performed 3 days apart. There was no difference in either the early pulmonary hypertension or the later increased permeability phase of endotoxin lung injury when comparing the paired studies, as measured by lymph flow and protein flux. Hemorrhagic shock, despite producing a transient increase in thromboxane and pulmonary leukocyte sequestration, does not accentuate the lung injury of endotoxin if the shock state is adequately resuscitated.     

Changes and significance of chemiluminescence of polymorphonuclear leukocytes in endotoxin-induced lung injury in conscious sheep]

The chemiluminescence (CL) of polymorphonuclear leukocytes and its relation with pulmonary microvascular permeability after endotoxin-induced lung injury in conscious sheep with lung lymph fistula were observed. Four hours after the injury the CL of PMNs increased from 0.27 cpm/PMN of baseline to 0.69 cpm/PMN (P < 0.05). The increment of the CL had positive correlation with the increment of lung lymph flow or permeability index (r = 0.632 0.638 P < 0.05), suggesting that the increase of pulmonary microvascular permeability after the endotoxin injury had relation with the increase of the respiratory tract of PMNs.     

Increased sheep lung vascular permeability caused by Escherichia coli endotoxin.

We infused Escherichia coli endotoxin, 0.07-1.33 microgram/kg, intravenously into chronically instrumented unanesthetized sheep and measured pulmonary arterial and left atrial pressures, lung lymph flow, lymph and blood plasma protein concentrations, and arterial blood gases. Endotoxin caused a biphasic reaction: an early phase of pulmonary hypertension and a long late phase of steady state increased pulmonary vascular permeability during which pulmonary arterial and left atrial pressures were not increased significantly and lung lymph flow was 5 times the baseline value. Lymph: plasma total protein concentration ratio during the late phase (0.76 +/- 0.04) was significantly (P less than 0.05) higher than during baseline (0.66 +/- 0.03). The lymph response was reproducible. Lung lymph clearance of endogenous proteins with molecular radii (r) 35.5 to 96 A was increased during the steady state late phase of the reaction, but, as during baseline, clearance decreased as r increased. The endotoxin reaction was similar to the reaction to infusing whole Pseudomonas bacteria, except that endotoxin had less effect on pressures during the steady state response and caused a relatively larger increase in lymph clearance of large proteins. We conclude that E. coli endotoxin in sheep causes a long period of increased lung vascular permeability and may have a greater effect on large solute pathways across microvessels than do Pseudomonas bacteria.     

Dibutyryl cyclic AMP attenuates lung responses induced by endotoxin in conscious sheep.

Dibutyryl cyclic AMP (DBcAMP) could inhibit the production of prostanoids and modulate the pulmonary vascular responses induced by endotoxin. Diffuse lung injury after endotoxemia in sheep is accompanied by the production of prostanoids and an increase in endothelial permeability. To determine whether exogenous DBcAMP could prevent the endotoxin responses, we measured pulmonary hemodynamics, gas exchange, and lung lymph responses to an intravenous infusion of Escherichia coli endotoxin (1.0 micrograms/kg over 30 min) in unanesthetized sheep in the presence and absence of DBcAMP (30 micrograms/kg/min) infused intravenously for 6 h beginning 1 h before endotoxin infusion or for 4.5 h after 30 min of treatment with endotoxin infusion. We also measured circulating leukocytes and lung lymph and plasma concentrations of thromboxane B2 (TXB2) and prostacyclin (6-keto-PGF1 alpha) metabolites by radioimmunoassay. DBcAMP infusion before endotoxin infusion decreased endotoxin-induced pulmonary hypertension and hypoxemia and markedly attenuated the increased lung lymph flow and lymph protein clearance. DBcAMP after endotoxin only attenuated the increased lung lymph flow and lymph protein clearance. DBcAMP treatment both before and after endotoxin infusion blocked endotoxin-induced increases in lung lymph and plasma TXB2 and 6-keto-PGF1 alpha. DBcAMP did not affect the number of circulating leukocytes. Although DBcAMP alone did not affect the pulmonary and systemic hemodynamics and lung lymph balance, the potential that DBcAMP directly modulates the pulmonary vascular responses to endotoxin as a vasodilator could be expected. We conclude that DBcAMP infusion attenuates lung dysfunction caused by endotoxemia, possibly by preventing prostanoid release and modulating the pulmonary vascular responses.     

TNF- and LPS-induced changes of lung vascular permeability: studies in unanesthetised sheep

There have been reports that the administration of TNF produces many of the cardiopulmonary changes seen with endotoxin (LPS). We asked whether all of LPS effects can be mimicked by TNF infusion. The effects of infusion of (human recombinant) TNF (50 micrograms/kg/30 min) and LPS (3 micrograms/kg/30 min) on permeability characteristics of sheep lungs were compared. Thirteen sheep were chronically instrumented for cardiopulmonary studies including lung lymph data. Infusion of LPS and TNF result in an increase of body temperature and lung lymph protein clearance (measured as the product of Lymph Flow and Lymph/Plasma Protein Ratio). The two responses had entirely different time courses. This might be related to the significantly longer period of pulmonary hypertension (MPAP) with LPS, which was only transient with TNF. Similar differences in plasma levels of thromboxane B2 (TxB2) were also noted. There were also differences in leukocyte kinetics, arterial blood gases, and plasma lactate levels. This indicates that although TNF infusion results in an increased pulmonary microvascular permeability, this event occurs without concomitant changes in thromboxane metabolism and exhibits time characteristics somewhat different from endotoxin.     

Relationship of increased lung serotonin levels to endotoxin-induced pulmonary hypertension in sheep. Effect of a serotonin antagonist

A persistent pulmonary artery hypertension, increased airways resistance, increased vascular permeability to protein, and hypoxia are characteristic of sepsis-induced ARDS in humans and are present in the late phase injury response seen in sheep after endotoxin. Our purpose was to determine the role of serotonin, 5-HT, in the steady-state pulmonary hypertension and decreased arterial oxygen tension seen beginning approximately 3 h after Escherichia coli endotoxin injury (2 micrograms/kg) in the adult sheep. Plasma 5-HT levels remained constant, whereas lung lymph values increased from a baseline of 60 +/- 40 to 180 +/- 70 and 270 +/- 90 ng/ml at 1-h and at 3- to 5-h periods, respectively, after endotoxin. Platelet count decreased significantly only at the 3-h time period. Ketanserin, a 5-HT antagonist, was infused (0.15 mg/kg/h) in 7 sheep during endotoxin injury. The degree of early pulmonary hypertension and hypoxia was not affected by ketanserin. Mean values for pulmonary artery pressure and arterial oxygen tension were 40 +/- 8 mmHg and 70 +/- 8 torr for endotoxin alone and 38 +/- 7 mmHg and 72 +/- 7 torr for the ketanserin group. Steady-state, protein-rich pulmonary perfusion was also not altered, being increased 3-fold in both groups. Pulmonary hypertension and hypoxia were significantly attenuated, however, at the 3- to 5-h period with ketanserin, compared with endotoxin alone, the pulmonary artery pressure decreasing from 29 +/- 5 to 22 +/- 4 mmHg and the PaO2 increasing from 75 +/- 4 to 83 +/- 5 torr.(ABSTRACT TRUNCATED AT 250 WORDS)     

Prostaglandin E1 prevents increased lung microvascular permeability during intravascular complement activation in sheep

Prostaglandin E1 (PGE1) inhibits a variety of functions of activated neutrophils including respiratory burst, release of leukotriene B4, and adherence to endothelial cells. To determine if PGE, alters the pathophysiology of complement-induced lung vascular injury, experiments were conducted in anesthetized sheep with lung lymph fistulas given a 1-hour infusion of zymosan-activated plasma. PGE1 (30 ng/min/kg) or its saline vehicle was infused intravenously for 90 minutes beginning 30 minutes before the infusion of activated plasma. PGE1 had no effect on leukocyte count, the initial hypoxemia and thromboxane A2 release, or the development of acute pulmonary hypertension. However, PGE1 prevented steady-state increases in lung lymph flow that in vehicle-treated sheep signaled an increase in lung microvascular permeability. Furthermore, extraction of PGE1 by pulmonary endothelial cells was unaffected by the infusion of activated plasma. We propose that PGE1 prevented the increase in lung vascular permeability by inhibiting adherence of activated neutrophils to endothelial cells.     

Recombinant-human superoxide dismutase attenuates endotoxin-induced lung injury in awake sheep.

Oxygen radicals have been implicated in the pathogenesis of acute lung injury associated with clinical and experimental sepsis. With the use of endotoxin infusion as an in vivo model of sepsis we studied the effect of recombinant-human superoxide dismutase (r-hSOD; 4,200 U/mg), an enzyme that catalyzes the dismutation of superoxide anion, on both the physiologic and biochemical lung changes in awake sheep. Sheep (n = 11) were prepared for chronic measurement of pulmonary hemodynamics and lung fluid balance. Paired experiments were performed in seven of the animals in which they received either endotoxin (1 microgram/kg) alone or in combination with r-hSOD in random order. An additional four sheep received r-hSOD without the lipopolysaccharide. Intravenous infusion of r-hSOD (a loading dose of 12,600 U/kg followed by a maintenance dose of 14,700 U/kg/h for 7 h) resulted in substantial SOD activity, measured by electron spin resonance spectrometry, both in plasma and in lung lymph, and attenuated the expected changes in pulmonary arterial pressure and lung lymph flow after administration of endotoxin. When administered without endotoxin, r-hSOD produced no perceptible change in pulmonary hemodynamics and lung fluid balance. These data suggest that superoxide anion plays an important role in endotoxin-induced lung injury in sheep.     

The effect of glucagon on the pulmonary transvascular fluid filtration rate

Administration of glucagon has been shown to decrease pulmonary vascular resistance, but its primary site of action is undetermined. Whether this is on the arterial or venous side of the capillary would be reflected in the microvascular hydrostatic pressure. We used the pulmonary flow of lymph, a sensitive index of the transvascular fluid filtration rate, to monitor the microvascular hydrostatic pressure. Eight unanesthetized sheep with a surgically created long-term fistula for monitoring pulmonary lymph were given a 3-mg bolus of glucagon after a baseline period. We found no change in pulmonary arterial or left atrial pressures but noted a significant increase in cardiac output and a decrease in pulmonary resistance. The flow of pulmonary lymph increased by 50 percent for 30 minutes after administration of glucagon, and the protein content of the lymph decreased by 15 percent, indicating a large increase in the microvascular hydrostatic pressure. From these data, we calculated a decrease in arterial resistance from 60 percent to 30 percent of the total and, subsequently, an increase of 6 cm H2O in the microvascular hydrostatic pressure. Administration of glucagon, therefore, decreases the arterial resistance while increasing microvascular pressure in the process.     

Effect of infusing histamine into pulmonary or bronchial artery on sheep pulmonary fluid balance

We determined the dose-response curve for the effect of histamine on pulmonary fluid and protein exchange when histamine was infused directly into the bronchial artery or pulmonary circulation in anesthetized sheep with a right thoracotomy and an acutely prepared lung lymph fistula. We measured pulmonary hemodynamics, lung lymph flow, and lymph-to-plasma protein concentration ratios in 14 sheep. We infused histamine at constant rates of 0.1, 1, 3, and 10 micrograms/kg . min until a steady state was achieved at each level for 1 h. In the bronchial circulation the lowest effective dose of histamine was 1 microgram/kg . min, which increased lymph flow at constant ratio of lymph to plasma protein concentration. Lymph flow increased further as the dose of infused histamine increased. Infusions into the pulmonary circulation had similar effects, but in addition, pulmonary vascular resistance decreased. The dose-response curves for the effects of histamine on lung lymph and protein flow were identical for the 2 routes of infusion. The effect of histamine was to increase lung microvascular permeability modestly. We conclude that the site of continued action of histamine is probably in the pulmonary microcirculation, because for any given infusion rate, the local concentration of histamine in the bronchial microcirculation must be much greater than that in the pulmonary microcirculation because the ratio of blood flows is so different.     

A 21-aminosteroid, U-74006F, attenuates endotoxin-induced lung injury in awake sheep

The purpose of the present study was to examine the efficacy of U-74006F, a 21-aminosteroid, on lung dysfunction induced by endotoxaemia in awake sheep with lung lymph fistula and haemodynamic monitoring. We measured pulmonary haemodynamics, lung lymph balance, circulating leucocyte count, arterial blood gas tensions, and levels of thromboxane (Tx) B2 and 6-keto-prostaglandin (PG) F1 alpha in plasma and lung lymph. We performed two experiments. In experiment 1 (n = 6), we intravenously infused Escherichia coli lipopolysaccharide endotoxin (1 microgram/kg) over 30 min and observed the parameters over 5 h. In experiment 2 (n = 6), we pretreated sheep with an intravenous bolus of U-74006F (2 mg/kg) 30 min before the infusion of endotoxin in the same manner of experiment 1, and continuously infused U-74006F (0.5 mg/kg per h) over 5 h after the bolus during the experiment. The U-74006F significantly suppressed the early pulmonary hypertension, the late increase in pulmonary permeability and the elevations of TxB2 and 6-keto-PGF1 alpha levels in plasma and lung lymph during the early period following endotoxaemia, although the compound did not change the time course of leucocytopenia and hypoxaemia. These findings suggest that the administration of U-74006F attenuates the lung dysfunction induced by endotoxaemia in awake sheep.     

Effects of alveolar hypoxia on lung fluid and protein transport in unanesthetized sheep.

To determine whether hypoxia directly affects pulmonary microvascular filtration of fluid or permeability to plasma proteins, we measured steady state lung lymph flow and protein transport in eight unanesthetized sheep breathing 10% O2 in N2 for 4 hours. We also studied three sheep breathing the same gas mixture for 48 hours. We surgically prepared the sheep to isolate and collect lung lymph and to measure average pulmonary arterial (Ppa) and left atrial (Pla) pressures. We placed a balloon catheter in the left atrium to elevate Pla. After recovery, the sheep breathed air through a tracheostomy for 2-4 hours, followed by 4 or 48 hours of hypoxia. In 13 4-hour studies, the average arterial PO2 fell from 97 to 38 torr; Ppa rose from 20 to 33 cm H2O; and lung lymph flow and lymph protein flow were unchanged. We also found that during 48-hour hypoxia, with a sustained elevation in Ppa and a decline in Pla, lymph flow and protein flow did not increase. In four sheep, we also raised Pla for 4 hours, followed by 4 hours of hypoxia with elevated Pla. Again, despite the added stress of elevated Pla, we found that lymph flow and lymph protein flow remained constant during hypoxia. We conclude that severe alveolar hypoxia, for 4 or 48 hours, alone or with increased pulmonary microvascular pressure, produced no change in lung fluid filtration or protein permeability, a finding supported by normal postmortem histology and extravascular lung water content.     

Role of subcutaneous tissue endotoxin in the production of prostanoid-induced lung injury: comparison with intravenous endotoxin response

Local injection of endotoxin into soft tissues of the flank results in hypoxia and pulmonary hypertension. Our purpose was to determine whether this was caused by tissue prostanoid production or production by the lung as is seen with endotoxemia. Twenty-six sheep were prepared with lung and flank tissue lymph fistulae. Thirteen sheep were given 2 micrograms/kg Escherichia coli endotoxin into the flank soft tissue, six of which were pretreated with ibuprofen, 12.5 mg/kg. Thirteen sheep were given intravenous endotoxin, 2 micrograms/kg, with six pretreated with ibuprofen. An early hypertensive phase was noted with both insults characterized by pulmonary hypertension, hypoxia, and increased lung lymph flow (QL). With subcutaneous tissue endotoxin, there was a significant increase in tissue lymph TxB2 and 6-keto-PGF1 alpha when compared to lung lymph and increased values in venous plasma compared to arterial plasma, indicating tissue to be the source. With intravenous endotoxin, lung lymph and aortic plasma levels were significantly higher than tissue lymph and venous plasma, respectively. The hypoxia, hypertension and increased prostanoids were prevented using ibuprofen. An increased lung permeability phase was noted with intravenous endotoxin but not with tissue endotoxin. As expected, this phase was not inhibited with ibuprofen and, therefore, not prostanoid-induced.     

Endotoxin and lung injury

Two things are certain: endotoxin has dramatic effects on the structure and function of the lungs in intact animals and also on isolated lung cells, and both the in vivo and in vitro effects of endotoxin are complex. In whole animals, endotoxin causes obvious and subtle effects on functions of both airways and the pulmonary circulation. These effects include diffuse lung inflammation and injury of pulmonary vascular endothelium. Endotoxin can also directly injure endothelial cells in vitro. In vivo, lung injury caused by endotoxin is at least partly dependent on the presence of granulocytes, and some evidence also suggests that both lymphocytes and macrophages may participate in the response either directly or by directing cell traffic. At least in the sheep preparation, platelets do not seem to play a major role in the lungs' response to endotoxemia. Although endotoxin can activate complement and activated complement infused into whole animals affects the lungs, it seems unlikely that complement activation alone is sufficient to explain the severe and prolonged lung injury caused by endotoxin. Cyclooxygenase metabolites of arachidonic acid appear to mediate both changes in lung mechanics and pulmonary vasoconstriction after endotoxemia. Lipoxygenase products may play a role in these responses as well as the inflammatory response and increases in vascular permeability, although evidence for these latter speculations is not firm. Lung cell injury caused by endotoxin probably is mediated at least in part by generation of free radicals. Inflammatory cells, especially neutrophils, are one source of these toxic oxygen species, but intracellular generation of free radicals within lung cells per se may also be stimulated by endotoxin and account for some of the lung injury. Likewise, inflammatory-cell-derived proteinases may mediate endotoxin-induced injury of lung cells and, as with chronic lung diseases, balance between proteinases and antiproteinases could be important. The fact that free radicals can inactivate antiproteinases, and antiproteinases can act as free radical scavengers, may suggest a complex relationship among the several possible mediators of toxicity. Cyclic nucleotide metabolism is affected in whole animals and isolated lung cells by endotoxin and these classic second messengers could be involved in the pathogenetic sequence, but exactly how is unclear. Chronic effects of endotoxin on the lungs may provide a pathogenetic link between acute lung injury and chronic changes in lung structure and function.(ABSTRACT TRUNCATED AT 400 WORDS)     

Failure of a protease inhibitor to affect the cardiopulmonary response to endotoxin when combined with a cyclooxygenase inhibitor

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.     

Influence of terbutaline on endotoxin-induced lung injury

The effects of the beta-2-receptor agonist terbutaline on central hemodynamics, gas exchange, and platelet and leukocyte counts during 3 hr of endotoxin shock were studied. Ten sheep were anesthetized and ventilated without positive end-expiratory pressure (PEEP). After 1 hr of stabilization (t = -30), five animals (Group T) received intravenous (i.v.) infusion of terbutaline, 20 micrograms/kg/hr for 3.5 hr, whereas the other five received no drug treatment and served as controls (Group C). Thirty min later (t = 0) all animals received Escherichia coli endotoxin 10 micrograms/kg by i.v. infusion over 15 min. The terbutaline infusion increased the heart rate (HR) initially by 30% and the cardiac index (CI) by 50%, whereas mean arterial pressure (MAP), pulmonary artery pressure (PAP), and gas exchange remained unchanged. Terbutaline pretreatment did not prevent the pulmonary hypertension that characteristically occurs after endotoxin injection, nor did it decrease the initial fall in platelet count, leukocyte count, arterial oxygen tension (PaO2), or respiratory compliance (t = 30). However, during the permeability phase (after 120-180 min), there was a significant improvement in MAP, PAP, respiratory compliance, PaO2, and arterial pH in the animals treated with terbutaline as compared with the control animals. Also, the wet weight to dry weight ratio of the lungs from animals receiving terbutaline was significantly lower than in controls. It was concluded that terbutaline does not influence the hypertension phase during endotoxin shock, but it may decrease pulmonary microvascular leakage during the permeability phase.