Effect of endotoxin on lung fluid balance in unanesthetized sheep

A Modern View of the Interstitial Space in Health and Disease

Increases in the volume of the interstitial space are readily recognized clinically as interstitial edema formation in the loose connective tissue of skin, mucosa, and lung. However, the contents and the hydrostatic pressure of this interstitial fluid can be very difficult to determine even in experimental settings. These difficulties have long obscured what we are beginning to appreciate is a dynamic milieu that is subject to both intrinsic and extrinsic regulation. This review examines current concepts regarding regulation of interstitial volume, pressure, and flow and utilizes that background to address three major topics of interest that impact IV fluid administration. The first of these started with the discovery that excess dietary salt can be stored non-osmotically in the interstitial space with minimal impact on vascular volume and pressures. This led to the hypothesis that, along with the kidney, the interstitial space plays an active role in the long-term regulation of blood pressure. Second, it now appears that hypovolemic shock leads to systemic inflammatory response syndrome principally through the entry of digestive enzymes into the intestinal interstitial space and the subsequent progression of enzymes and inflammatory agents through the mesenteric lymphatic system to the general circulation. Lastly, current evidence strongly supports the non-intuitive view that the primary factor leading to inflammatory edema formation is a decrease in interstitial hydrostatic pressure that dramatically increases microvascular filtration.

Interstitial Fluid and Lymph Formation and Transport: Physiological Regulation and Roles in Inflammation and Cancer

The interstitium describes the fluid, proteins, solutes, and the extracellular matrix (ECM) that comprise the cellular microenvironment in tissues. Its alterations are fundamental to changes in cell function in inflammation, pathogenesis, and cancer. Interstitial fluid (IF) is created by transcapillary filtration and cleared by lymphatic vessels. Herein we discuss the biophysical, biomechanical, and functional implications of IF in normal and pathological tissue states from both fluid balance and cell function perspectives. We also discuss analysis methods to access IF, which enables quantification of the cellular microenvironment; such methods have demonstrated, for example, that there can be dramatic gradients from tissue to plasma during inflammation and that tumor IF is hypoxic and acidic compared with subcutaneous IF and plasma. Accumulated recent data show that IF and its convection through the interstitium and delivery to the lymph nodes have many and diverse biological effects, including in ECM reorganization, cell migration, and capillary morphogenesis as well as in immunity and peripheral tolerance. This review integrates the biophysical, biomechanical, and biological aspects of interstitial and lymph fluid and its transport in tissue physiology, pathophysiology, and immune regulation.

Pathophysiology of tissue fluid accumulation in inflammation

The potential role of extravascular factors for the local as well as systemic response to an inflammatory stimulus is addressed here in light of recent data from the trachea, serving as a surrogate for lower airways, and spleen, because of its role in the immune response and fluid volume regulation. From analysis of interstitial fluid from trachea it is apparent that the colloid osmotic pressure is high relative to plasma, suggesting a significant buffering capacity against oedema formation, and also that there is a significant local production of proinflammatory mediators to a systemic inflammatory stimulus. Inflammatory stimuli may furthermore result in a rapid reduction in interstitial fluid pressure, thus leading to increased filtration and oedema formation. Knowledge regarding the fluid phase within the spleen microenvironment can be gathered via analysis of drained lymph. During a septic response induced by lipopolysaccharide injection, the spleen contributes significantly to the production of pro- and anti-inflammatory cytokines, and may induce protracted inflammation because of a dominant role in IL-6 production. Significant amounts of immune cells exit via lymph, and acquire specific activation signatures having been exposed to the spleen microenvironment. Although often overlooked, extravascular or interstitial factors may therefore contribute significantly to the inflammatory process and thus the ensuing oedema associated with inflammation.

Cytokines are produced locally by myocytes in rat skeletal muscle during endotoxemia

Cytokines act as chemical mediators during the inflammatory process. Measurements of cytokine levels in tissue have previously been performed in homogenized tissue, but the true concentrations in native interstitial fluid (ISF), i.e., the compartment where cytokines exert their biologically active role, have remained unknown. The role of skeletal muscle myocytes as a source for cytokines during endotoxemia was explored by collecting muscle ISF using a wick method, and the levels of 14 cytokines in ISF and plasma were related to the corresponding changes in mRNA levels to reveal any potential discrepancies between gene expression and protein release of cytokines to ISF. The majority of investigated cytokines were elevated in muscle ISF during endotoxemia, and an analysis of cytokine mRNA levels revealed consistency between gene expression and protein release. The elevated cytokine level in ISF, in addition to elevated gene expression in muscle, indicated a significant local production and release of several proinflammatory cytokines and chemokines within skeletal muscle tissue during endotoxemia. Immunohistochemistry revealed that myocytes constituted a significant source of IL-1beta and TNF-alpha production during endotoxemia, whereas the contribution from inflammatory cells i.e., leukocytes, was found to be less significant. Muscle cells apparently constitute an important source of several different cytokines during endotoxemia, governing the level in the muscle microenvironment, and are likely to contribute significantly to cytokine levels in plasma.

Isolation of rat trachea interstitial fluid and demonstration of local cytokine production in lipopolysaccharide-induced systemic inflammation

Access to interstitial fluid from trachea is important for understanding tracheal microcirculation and pathophysiology. We tested whether a centrifugation method could be applied to isolate this fluid in rats by exposing excised trachea to G forces up to 609 g. The ratio between the concentration of the equilibrated extracellular tracer 51Cr-labeled EDTA in fluid isolated at 239 g and plasma averaged 0.94 +/- 0.03 (n = 14), suggesting that contamination from the intracellular fluid phase was negligible. The protein pattern of the isolated fluid resembled plasma closely and had a protein concentration 83% of that in plasma. The colloid osmotic pressure in the centrifugate in controls (n = 5) was 18.8 +/- 0.6 mmHg with a corresponding pressure in plasma of 22 +/- 1.5 mmHg, whereas after overhydration (n = 5) these pressures fell to 9.8 +/- 0.4 and 11.9 +/- 0.4 mmHg, respectively. We measured inflammatory cytokine concentration in serum, interstitial fluid, and bronchoalveolar lavage fluid in LPS-induced inflammation. In control animals, low levels of IL-1 beta, IL-6, and TNF-alpha in serum, trachea interstitial fluid, and bronchoalveolar lavage fluid were detected. LPS resulted in a significantly higher concentration in IL-1 beta and IL-6 in interstitial fluid than in serum, showing a local production. To conclude, we have shown that interstitial fluid can be isolated from trachea by centrifugation and that trachea interstitial fluid has a high protein concentration and colloid osmotic pressure relative to plasma. Trachea interstitial fluid may also reflect lower airways and thus be of importance for understanding, e.g., inflammatory-induced airway obstruction.

Edemagenic gain and interstitial fluid volume regulation

Under physiological conditions, interstitial fluid volume is tightly regulated by balancing microvascular filtration and lymphatic return to the central venous circulation. Even though microvascular filtration and lymphatic return are governed by conservation of mass, their interaction can result in exceedingly complex behavior. Without making simplifying assumptions, investigators must solve the fluid balance equations numerically, which limits the generality of the results. We thus made critical simplifying assumptions to develop a simple solution to the standard fluid balance equations that is expressed as an algebraic formula. Using a classical approach to describe systems with negative feedback, we formulated our solution as a "gain" relating the change in interstitial fluid volume to a change in effective microvascular driving pressure. The resulting "edemagenic gain" is a function of microvascular filtration coefficient (K(f)), effective lymphatic resistance (R(L)), and interstitial compliance (C). This formulation suggests two types of gain: "multivariate" dependent on C, R(L), and K(f), and "compliance-dominated" approximately equal to C. The latter forms a basis of a novel method to estimate C without measuring interstitial fluid pressure. Data from ovine experiments illustrate how edemagenic gain is altered with pulmonary edema induced by venous hypertension, histamine, and endotoxin. Reformulation of the classical equations governing fluid balance in terms of edemagenic gain thus yields new insight into the factors affecting an organ's susceptibility to edema.

An increase of abdominal pressure increases pulmonary edema in oleic acid-induced lung injury

Increased abdominal pressure is common in intensive care unit patients. To investigate its impact on respiration and hemodynamics we applied intraabdominal pressure (aIAP) of 0 and 20 cm H(2)O (pneumoperitoneum) in seven pigs. The whole-lung computed tomography scan and a complete set of respiratory and hemodynamics variables were recorded both in healthy lung and after oleic acid (OA) injury. In healthy lung, aIAP 20 cm H(2)O significantly lowered the gas content, leaving the tissue content unchanged. In OA-injured lung at aIAP 0 cm H(2)O, the gas content significantly decreased compared with healthy lung. The excess tissue mass (edema) amounted to 30 +/- 24% of the original tissue weight (455 +/- 80 g). The edema was primarily distributed in the base regions and was not gravity dependent. Heart volume, central venous, pulmonary artery, wedge, and systemic arterial pressures significantly increased. At aIAP 20 cm H(2)O in OA-injured lung, the central venous and pulmonary artery pressures further increased. The gas content further decreased, and the excess tissue mass rose up to 103 +/- 37% (tissue weight 905 +/- 134 g), with homogeneous distribution along the cephalocaudal and sternovertebral axis. We conclude that in OA-injured lung, the increase of IAP increases the amount of edema.

Inhaled nitric oxide reduces lung fluid filtration after endotoxin in awake sheep

We studied the effect on lung fluid filtration of 37.6 ppm inhaled nitric oxide (NO) imposed for 1 h 2.5 h after endotoxin in seven awake sheep, with seven control subjects. The effects of NO on the longitudinal distribution of pulmonary vascular resistance (PVR) before and after endotoxin were specifically addressed in six sheep. Following endotoxin, sheep developed respiratory distress; PaO2, the alveolar-arterial oxygen tension difference (AaPO2) and venous admixture (Q S/Q T) changed significantly, as did the pulmonary artery pressure (Ppa), PVR, and lung lymph flow (Q L). Inhaled NO reduced Ppa and PVR by 50%; Q L decreased from 7.8 +/- 0.34 ml/15 min to 4.7 +/- 0.80 ml/15 min (mean +/- SEM), and lymph protein clearance from 4.9 +/- 0.18 ml/15 min to 3.6 +/- 0.75 ml/15 min. Lymph/plasma protein concentration ratio (L/P) increased from 0.63 +/- 0.016 to 0.72 +/- 0.006, concomitant with the decrease in Q L. The L/P - Q L relationships shifted from left, at baseline, to the right during endotoxemia, as did the permeability surface product (PS) isolines. The rightward shift was significantly less in the NO group. Inhaled NO significantly improved PaO2, AaPO2, and Q S/Q T, reduced the increase in pulmonary microwedge pressure back to baseline and decreased upstream and downstream PVR at 3.0 through 4. 0 h. We conclude that, in sheep, inhaled NO reduces lung fluid filtration by decreasing microvascular pressure and apparently also by declining the enhanced microvascular permeability during the late phase of endotoxemia.

Evaluation of pulse oximetry in horses surgically treated for colic

All 43 horses anaesthetised for colic surgery were premedicated with xylazine or diazepam. Anaesthesia was induced with guaifenesin and ketamine, horses were placed in dorsal recumbency and anaesthesia was maintained with isoflurane in oxygen and mechanical ventilation. Haemoglobin saturation readings (SpO2) were taken with a pulse oximeter and compared with computed haemoglobin saturation (SaO2) from arterial blood samples. Readings were taken over a range of SaO2 of 78-100%, mean arterial blood pressure ranged from 24 to 108 mmHg and PaO2 ranged from 53 to 490 mmHg. Analysis of 107 readings showed that SpO2 values predicted SaO2 but time, blood pressure and individual horse did not. Correlation coefficients between SpO2 and SaO2 were 0.85 for all values and 0.88 for values at 30 min. Values for bias and precision were calculated for all SpO2 values and for readings separated into 3 saturation groups: normal, low normal, and abnormal. The pulse oximeter tended to underestimate SaO2 at all times, and was less precise as the saturation decreased.

Clinical detection of LPS and animal models of endotoxemia

The interest in the study of endotoxemia in the clinical area has increased recently as a result of a) improved and simplified endotoxin determination e.g. chromogenic-kinetic microplate methods (also an improved blood sampling tool is available), b) incidence of sepsis has increased due to improvement in early (e.g. posttraumatic) survival, c) interest in and good evidence for gut translocation as a source of endotoxemia, d) agents have developed, which can antagonize endotoxins. There is evidence that patients with positive endotoxin test in the ICU have a higher incidence of organ failure. To study the pathophysiological consequences of endotoxemia and possible ways of intervention animal models are necessary. The choice of the experimental setting depends on the aim of the study e.g. whether prolonged observation is necessary in survival studies or whether hemodynamic variables have to be measured or whether therapeutic agents only crossreact with primates. Since LPS levels are quite low in clinical studies, an important factor for selection of a relevant animal might be LPS sensitivity, or the use of additional sensitization techniques e.g. galactosamine. Another important aspect in this context is whether LPS is given as bolus or infused up to several days. In this review the dose, time, and route of LPS administration is also discussed. For screening purposes rodents are usually used, or sometimes rabbits due to their higher LPS sensitivity. Another very sensitive animal model is the sheep, which can be chronically instrumented and as a specialty allows lung lymph drainage and thus studies of LPS effects on pulmonary permeability. Pigs are used for hemodynamic studies and often in therapeutical studies if species-specificity of the drug tested is not important, in cases where a large animal is necessary. Finally the non-human primates offer a number of advantages due to human-like physiology, due to the cross-reactivity of human assay systems and accordingly also cross-reactivity of human therapeutic agents. While the chimpanzee also shares the LPS sensitivity of humans, baboons are insensitive like rodents. Thus each model serves to provide some useful purpose and the selection must be made to meet the requirements of the specific questions to be asked, with special emphasis of the chosen endotoxin model on relevance for the human sepsis state.

Lymph and blood flow responses in central airways

The lymphatic drainage of the lung has been used as a quantitation of pulmonary microvascular fluid flux in normal animals and after various forms of injury. This review supports the importance of the bronchial microvasculature in the formation of lung lymph. Proof that the lymph drainage of the lung comes from the pulmonary circuit has been based on the finding of an elevation of lymph flow when the pulmonary venous pressure is elevated. This proof is wanting since recent work demonstrates that the venous drainage of the intrapulmonary bronchi flows into the pulmonary vascular system at the precapillary level. The administration of endotoxin induces an elevation of lung lymph. The bronchial circuit may play a role in this response since it is likewise exposed to the high pulmonary pressures induced by endotoxin, and there is evidence that ischemia/reperfusion injury to the airway occurs with endotoxin administration. After acute lung injury from smoke inhalation, lung lymph flow is markedly elevated. The lymph drainage from the airway may play an important role in this response. Bronchial blood flow is markedly increased after inhalation injury and there is airway edema. The increases in lung lymph flow and extravascular lung water are markedly reduced by occlusion of the bronchial artery. These data support the need for additional study of the role of the bronchial circulation in the formation of lung lymph.

Pneumonia complicating abdominal sepsis: an experimental model of hematogenous contamination of the lung

Pulmonary infection complicating intra-abdominal sepsis is a major clinical problem. An experimental model for intra-abdominal sepsis was created with implantation of gelatin capsules, containing 3 x 10(8) cfu E. coli strain no. 2554, in the peritoneal cavity of 20 rats (10 animals received and 10 did not receive antibiotic therapy with ceftriaxone) in order to verify the role of the primary site of infection in the pathogenesis of pneumonia. Ten rats were sacrificed to determine the relative pulmonary weight and 10 were submitted to simple laparotomy and insertion of a germ-free capsule (sham-operated group). In this group of animals there was only one death (10%). All the rats that received antibiotic therapy survived until sacrifice while all the rats that did not receive ceftriaxone died, 7 within the 2nd and 3 on the 6th postoperative day. Pneumonia and peritonitis developed only in the animals that did not receive ceftriaxone. Bacteriological findings of material obtained from peritoneal and pleural cavities revealed the same strain of E. coli used for the experiment, suggesting that bacteria involved in the pleuro-pulmonary infections may originate in the primary site of infection and that antibiotic therapy started at the moment of contamination, can prevent this major complication.

Differential responses of the endothelial and epithelial barriers of the lung in sheep to Escherichia coli endotoxin

Although intravenous Escherichia coli endotoxin has been used extensively in experimental studies to increase lung endothelial permeability, the effect of E. coli endotoxin on lung epithelial permeability has not been well studied. To examine this issue in sheep, bidirectional movement of protein across the lung epithelial barrier was studied by labeling the vascular space with 131I-albumin and by instilling 3 ml/kg of an isosmolar protein solution with 125I-albumin into the alveoli. E. coli endotoxin was administered according to one of three protocols: intravenous alone (5-500 micrograms/kg), intravenous (5 micrograms/kg) plus low-dose alveolar endotoxin (10 micrograms/kg), and high-dose alveolar endotoxin alone (50-100 micrograms/kg). Alveolar liquid clearance was estimated based on the concentration of the instilled native protein. Sheep were studied for either 4 or 24 h. Although intravenous E. coli endotoxin produced a marked increase in transvascular protein flux and interstitial pulmonary edema, there was no effect on the clearance of either the vascular (131I-albumin) or the alveolar (125I-albumin) protein tracer across the epithelial barrier. High-dose alveolar E. coli endotoxin caused a 10-fold increase in the number of leukocytes, particularly neutrophils, that accumulated in the air spaces. In spite of the marked chemotactic effect of alveolar endotoxin, there was no change in the permeability of the epithelial barrier to the vascular or alveolar protein tracers. Also, alveolar epithelial liquid clearance was normal. Morphologic studies confirmed that the alveolar epithelial barrier was not injured by either intravenous or alveolar E. coli endotoxin. Thus, the alveolar epithelium in sheep is significantly more resistant than the lung endothelium to the injurious effects of E. coli endotoxin.

Recombinant tumor necrosis factor increases pulmonary vascular permeability independent of neutrophils

We studied the effects of intravenous infusion of recombinant human tumor necrosis factor type alpha (rTNF-alpha; 12 micrograms/kg) on lung fluid balance in sheep prepared with chronic lung lymph fistulas. The role of neutrophils was examined in sheep made neutropenic with hydroxyurea (200 mg/kg for 4 or 5 days) before receiving rTNF-alpha. Infusion of rTNF-alpha resulted in respiratory distress and 3-fold increases in pulmonary arterial pressure and pulmonary vascular resistance within 15 min, indicating intense pulmonary vasoconstriction. Pulmonary lymph flow (i.e., net transvascular fluid filtration rate) and transvascular protein clearance rate (a measure of vascular permeability to protein) increased 2-fold within 30 min. The increased permeability was associated with leukopenia and neutropenia. The pulmonary hypertension and vasoconstriction subsided but fluid filtration and vascular permeability continued to increase. Sheep made neutropenic had similar increases in pulmonary transvascular fluid filtration and vascular permeability. rTNF-alpha also produced concentration-dependent increases in permeability of 125I-labeled albumin across ovine endothelial cell monolayers in the absence of neutrophils or other inflammatory mediators. The results indicate that rTNF-alpha increases pulmonary vascular permeability to protein by an effect on the endothelium.

Effect of antibody-mediated neutropenia on the cardiopulmonary response to endotoxemia

An ovine anti-neutrophil antibody has been produced by immunizing rabbits with purified sheep neutrophils. Serial intraarterial infusions of anti-neutrophil antibody in awake instrumented sheep produced selective and profound neutropenia. Intravascular infusion of endotoxin (Escherichia coli, 1.5 micrograms/kg/30 min) resulted in significant and equivalent increases in pulmonary artery pressure, peripheral vascular resistance, and protein-rich pulmonary lymph flow in an endotoxin group (n = 9) and a depletion + endotoxin group (n = 4). Changes in cardiopulmonary parameters were most pronounced 2 to 8 hr after endotoxin administration in both groups. Cardiac index (CI) showed a precipitous and transient fall in both experimental groups at 0.5 to 1 hr after endotoxin infusion; however, by 8 hr CI rose significantly in the endotoxin group, while it remained unchanged in the depletion + endotoxin group. A significant rise in the peripheral neutrophil count was associated with the increase in CI in the endotoxin group. Plasma and pulmonary lymph levels of thromboxane-B2 were unchanged during the depletion period with a significant increase 1 hr after endotoxin infusion. In this study questions arise regarding the exclusive role of circulating neutrophils in the microvascular permeability changes seen in sepsis-mediated adult respiratory distress syndrome.

Effect of catalase on endotoxin-induced acute lung injury in unanesthetized sheep

Administration of endotoxin intravenously to unanesthetized sheep causes an acute lung injury characterized by increased microvascular barrier permeability and subsequent pulmonary edema. Endotoxin-induced sheep lung injury can be attenuated by leukocyte depletion, and may be mediated by toxic metabolites of oxygen. We studied effects of administering catalase, which catalyzes conversion of hydrogen peroxide to oxygen and water, to sheep subsequently infused with endotoxin to test the hypothesis that hydrogen peroxide plays a role in the pathogenesis of lung injury. We found that infusions of endotoxin (1 microgram/kg) into untreated sheep caused the expected biphasic response, a transient, early, marked pulmonary arterial hypertension followed by a prolonged increase in protein-rich lung lymph flow characteristic of increased microvascular permeability filtration in the lungs. Intraperitoneal injections of catalase (50 mg/kg) prior to infusing endotoxin in these same sheep resulted in substantial catalase activity in plasma and in lung lymph, and attenuated the expected changes in pulmonary arterial pressure, lung lymph flow, and arterial leukocyte counts and oxygen tension after endotoxin infusions. Furthermore, mechanical elevation of hydrostatic pressure in the lungs of a catalase-treated sheep infused with endotoxin resulted in increased lung lymph flow with a decreased protein concentration, indicating that the microvascular barrier to fluid and protein was functionally intact. Administration of catalase that was inactivated by reaction with hydrogen peroxide in the presence of aminotriazole or administration of the catalase vehicle, thymol, had no effects on the sheep responses to endotoxin. We conclude that hydrogen peroxide plays a role in the pathogenesis of endotoxin-induced acute lung injury in sheep.

A model of the lung interstitial-lymphatic system

Our model of the pulmonary interstitial-lymphatic system is based on the assumption that the lung interstitial space can be divided into two compartments. The first compartment (C1) contains the terminal lymph vessels. Increases in the fluid pressure within this compartment, along with increased pressure generated by lymph vessel pumping, cause the lymph flow rate to increase. The lymph vessels run through the second compartment (C2) which we believe represents the perivascular spaces. Increases in the fluid volume of C2 cause the lymph vessels to dilate and this causes lymph vessel resistance to decrease. Normally the lymph flow rate equals the microvascular filtration rate so that lung fluid volume is constant. According to our model, increases in filtration rate cause fluid to collect in C1 and C2. The resulting increase in fluid pressure in C1, increased lymph vessel pumping, and the decrease in lymph vessel resistance in C2 cause lymph flow to increase. Eventually, the lymph flow rises to equal the filtration rate and lung fluid volume becomes constant again. The results of simulations with our model indicate that decreases in lymph vessel resistance are essential for lymph flow to increase substantially as edema develops.

Interference of the PAF-acether antagonist BN 52021 with endotoxin-induced hypotension in the guinea-pig

Because of the potential role of PAF-acether in the pathogenesis of endotoxin shock, we examined the preventive and curative effects of BN 52021, a new PAF-acether antagonist in guinea-pig challenged with S. Typhimurium endotoxin. A biphasic reduction of mean arterial pressure was elicited by i.v. endotoxin (300 micrograms/kg) in control animals, with a rapid drop of blood pressure (maximal decrease within 10 min), partial recovery at 20 min and a second gradual decrease after 30 min. Treatment with BN 52021 injected 15 min prior to endotoxin reduced the initial rapid drop of blood pressure from 38.5 +/- 5 mmHg in vehicle-treated controls (n = 15) to 17 +/- 3 mmHg (p less than 0.01) in animals treated with 1 mg/kg BN 52021(n = 10) and to 9.5 +/- 8 mmHg (p less than 0.01) in guinea-pigs treated with 6 mg/kg BN 52021 (n = 5). The early hypotensive phase was associated with severe thrombocytopenia-leukopenia; only the thrombocytopenia was reduced by BN 52021. The prolonged secondary phase of hypotension was reduced by BN 52021 pretreatment whereas a small increase of hematocrit persisted. The two phases of the arterial pressure profile during endotoxic shock were not observed in animals previously made thrombopenic by rabbit and anti-platelet serum and only the late hypotensive phase persisted. This late hypotension induced by endotoxin in thrombopenic animals was suppressed by BN 52021 pretreatment suggesting that BN 52021 may act via a platelet-independent mechanism. The intravenous injection of BN 52021 during the prolonged secondary phase of shock was followed by an immediate increase of the depressed blood pressure. This increase of blood pressure was dose-dependent, maximum at 6 mg/kg BN 52021, and observed in normal and thrombopenic animals. The interference of BN 52021 with endotoxin shock may be related to its PAF-acether antagonist properties and suggests that PAF-acether is an important participant in endotoxic shock.