Relation of blood-free to blood-inclusive postmortem lung water measurements in sheep

Our purpose was to see if the postmortem weight ratio of extravascular lung water to blood-free dry lung (blood-free ratio) was related to similar ratios in blood-inclusive lung and in blood. We developed linear regressions of blood-free ratio on ratios for blood-inclusive lung and blood together and for blood-inclusive lung alone for 73 sheep studied under 11 different protocols and for two subgroups of sheep, one with plasma space expansion and the other without expansion. The relation of ratios of blood-free to blood-inclusive lungs was different between the two subgroups. Although all regressions were highly correlated, the fits of the blood-free ratio on ratios for blood-inclusive lung and blood together were better than for blood-inclusive lung alone. The mean error of prediction of extravascular lung water for all sheep was significantly less for the regression of blood-free ratio on ratios for blood and blood-inclusive lung together (11 g) than for blood-inclusive lung alone (18 g). This study shows that weights of lung homogenate and blood samples before and after simple oven drying can be used to provide accurate inexpensive estimates of postmortem extravascular lung water.     

Evaluating flux of labeled albumin into the pulmonary interstitium of sheep lungs.

A noninvasive method was used to measure the movement of 131I-labeled albumin across the pulmonary microvascular barrier of a blood-perfused in situ sheep lung lymph preparation. After injection of labeled albumin into the blood, external measurements of gamma activity were taken for 2 h. The interstitial concentrations were calculated by applying the external activities and sampled lung lymph concentrations to a mass transport model. For the external activities and lymph activities to yield the same quantitative results, two modifications were necessary. First, lymph concentrations were corrected for transport delay from the lymphatic system. Second, externally detected radioactivity had to be corrected for the contribution of unbound nuclide. Application of a mathematical model to the data indicated the extravascular distribution volume for albumin was 79% of the pulmonary blood volume, and the extravascular distribution volume for radiolabeled iodide was 4.42 times greater than the pulmonary blood volume. The permeability-surface area product for iodide in the lung was estimated to be 0.274 ml.min-1.g blood-free dry lung wt-1. The transport delay in the lymphatic system was approximately 30-45 min and represented a volume of 1.44-2.80 ml.     

Spontaneous resolution of pulmonary edema caused by short periods of cyclic overinflation.

Mechanical ventilation with high or even moderate peak inspiratory pressure produces pulmonary permeability edema. Besides the level of overinflation, duration may affect both severity and type of edema. We studied the effect of 2 min of 35-mmHg peak pressure mechanical ventilation (HV) on microvascular permeability and deep lung fluid balance in rats. It resulted in increased extravascular lung water (+50%), bloodless dry lung weight (+25%), and albumin uptake in lungs (+450%). The increase in dry lung weight and albumin uptake compared with that of lung water suggested major permeability alterations. Ultrastructural examination showed the presence of numerous endothelial blebs. Epithelial lining fluid (ELF) volume, its potassium and protein concentrations, and cellular composition were assessed by bronchoalveolar lavage. There was an increase in ELF volume (+180%), a decrease in ELF potassium concentration (-50%), and an increase in ELF protein content (+76%). A few blood cells were recovered, suggesting the presence of a few large epithelial breaks. Some animals were allowed to recover for periods less than or equal to 180 min after HV. Extravascular lung water, dry lung weight, and albumin distribution space returned to control levels within 45 min. ELF volume diminished but remained larger than in controls, and ELF protein concentration increased probably because of alveolar fluid resorption. No further hemorrhage was observed. These results indicate that periods of HV as short as 2 min transiently alter microvascular permeability in rats.     

Vascular permeability and late radiation fibrosis in mouse lung

It has been suggested that fibrosis which develops after irradiation is caused by increases in vascular permeability. Plasma proteins leak into irradiated tissue where fibrinogen may be converted into fibrin which is gradually replaced by fibrous tissue. Vascular and fibrotic changes in mouse lung were investigated after X irradiation of the right hemithorax. Blood volume and accumulation of extravascular proteins were measured using indium (111In)-labeled red cells, iodinated (131I) albumin, and iodinated (125I) fibrinogen. Tracers were injected 1-47 weeks after irradiation and lungs were excised 24 or 96 hr later to determine radioactivity. The amount of collagen was estimated by measuring the hydroxyproline content. During the first few months after X rays, lung blood volume decreased to a plateau which depended on radiation dose (10-25 Gy). Small increases in extravascular albumin and fibrinogen occurred at 1-12 weeks after 10-25 Gy. Subsequently, protein returned to normal after 10 Gy, remained elevated after 15 Gy, and increased after 20 and 25 Gy. Hydroxyproline per gram of dry irradiated lung was increased at 18 weeks after 15-25 Gy. Subsequently it showed little change although both total hydroxyproline content and dry weight decreased after 20 and 25 Gy. Support for the hypothesis was that hydroxyproline per gram only increased after X-ray doses which caused marked extravasation of protein. There was no evidence, however, for deposition of 125I-fibrin or for a gradual increase in fibrosis corresponding to the prolonged excess of extravascular protein.     

Regional extravascular and interstitial lung water in normal dogs

We measured the regional distribution of pulmonary extravascular and interstitial water to examine the possibility that regional differences in microvascular pressure or tissue stress may cause regional differences in lung water. We placed chloralose-anesthetized dogs in an upright (n = 6) or supine (n = 7) position for 180 min. We injected 51Cr-labeled EDTA to equilibrate to the extracellular space and 125I-labeled albumin to equilibrate with plasma. At the end of the experiment, the lungs were removed, passively drained of blood, and inflated before rapid freezing. Lungs were divided into horizontal slices, and extravascular, interstitial, and plasma water, red cell volume, and dry lung weight were determined for each slice. We found that regional extravascular and interstitial water were constant throughout the lungs in both groups and that there were no significant differences between upright and supine dogs. There were no significant differences in hematocrit between slices. We conclude that gravity and body position have no measurable effect on either the total size of the extravascular and interstitial compartments or their regional distribution.     

Pulmonary angiotensin-converting enzyme substrate hydrolysis during exercise.

We examined exercise-induced changes in indicator-dilution estimates of the angiotensin-converting enzyme first-order kinetic parameter, the ratio of a normalized maximal enzymatic conversion rate to the Michaelis constant (Amax/Km), which, under stable enzymatic conditions, will vary with the pulmonary vascular surface area accessible to vascular substrate, the extravascular lung water (an index of the proportion of lung tissue perfused), and the central blood volume (from pulmonary trunk to aorta). Experiments were performed in 10 mongrel dogs at rest and through two increasing levels of treadmill exercise, with the use of two vascular space tracers (labeled erythrocytes and albumin), a water space tracer ([1,8-14C]-octanediol), and a vascular endothelium surface area marker, benzoyl-Phe-Gly-Pro ([3H]BPGP), which is a pharmacologically inactive angiotensin-converting enzyme substrate. The exercise-induced increase in cardiac output was accompanied by a linear increase in central blood volume, and dilutional extravascular lung water rapidly increased to an asymptotic proportion close to 100% of postmortem vascular lung water. There was an average 55% [3H]BPGP hydrolysis, which did not vary with flow, and the computed Amax/Km increased linearly with exercise. We conclude that exercise results in complete lung tissue recruitment and increases the pulmonary vascular surface area available for BPGP hydrolysis linearly with flow, so that pulmonary vascular recruitment continues after full tissue recruitment.     

Oleic acid-induced lung injury in rabbits: effect of fibrinogen depletion with Arvin

The role of fibrinogen in the evolution of the increased permeability after oleic acid-induced lung injury was studied in New Zealand White rabbits. Animals depleted of fibrinogen by treatment with Malayan pit viper venom were compared with untreated rabbits immediately and at 1 and 24 h after injury. The increased permeability to albumin and elevated extravascular lung water (EVLW) associated with lung injury returned to control values by 24 h in untreated animals. Fibrinogen-depleted animals had a higher mortality (10/25 vs. 2/17, P less than 0.02) and showed a greater immediate increase in permeability to albumin that returned to control values at 1 and 24 h after injury, as well as trends toward elevated blood-free dry lung weight and larger increases in EVLW that persisted for 24 h. These findings indicate that fibrinogen-related proteins play an important role in controlling the microvascular injury that is produced by oleic acid. However, when these proteins are depleted, other mechanisms partially control the leak at later stages of the repair process.     

Role of pulmonary blood flow in postpneumonectomy lung growth.

To study the influence of blood flow on postpneumonectomy lung growth, we banded the left caudal lobe pulmonary artery of eight ferrets in such a way that blood flow to the caudal lobe did not increase when the right lung was excised 1 wk later. The fraction of the cardiac output received by the right lung before pneumonectomy was therefore directed entirely to the left cranial lobe. Three weeks after pneumonectomy the weight, volume, and protein and DNA contents of the two lobes of the left lung were measured and compared with those of five unoperated animals and eight animals after right pneumonectomy alone. Although its perfusion did not increase after pneumonectomy, the left caudal lobe of banded animals participated in compensatory growth, increasing in weight and protein and DNA contents. Although the cranial lobe of banded animals received 25% more of the cardiac output than the same lobe in pneumonectomized animals, cranial lobe volume and protein and DNA contents in the two groups were similar. Caudal lobes were smaller in banded than in simple pneumonectomized animals and tended to contain less protein, whereas the cranial lobes tended to be heavier. We conclude that increased pulmonary perfusion is not necessary for compensatory lung growth in adult ferrets, but it may modify this response.     

Preventing mediastinal shift after pneumonectomy does not abolish physiological compensation.

To determine the role of mediastinal shift after pneumonectomy (PNX) on compensatory responses, we performed right PNX in adult dogs and replaced the resected lung with a custom-shaped inflatable silicone prosthesis. Prosthesis was inflated (Inf) to prevent mediastinal shift, or deflated (Def), allowing mediastinal shift to occur. Thoracic, lung air, and tissue volumes were measured by computerized tomography scan. Lung diffusing capacities for carbon monoxide (DL(CO)) and its components, membrane diffusing capacity for carbon monoxide (Dm(CO)) and capillary blood volume (Vc), were measured at rest and during exercise by a rebreathing technique. In the Inf group, lung air volume was significantly smaller than in Def group; however, the lung became elongated and expanded by 20% via caudal displacement of the left hemidiaphragm. Consequently, rib cage volume was similar, but total thoracic volume was higher in the Inf group. Extravascular septal tissue volume was not different between groups. At a given pulmonary blood flow, DL(CO) and Dm(CO) were significantly lower in the Inf group, but Vc was similar. In one dog, delayed mediastinal shift occurred 9 mo after PNX; both lung volume and DL(CO) progressively increased over the subsequent 3 mo. We conclude that preventing mediastinal shift after PNX impairs recruitment of diffusing capacity but does not abolish expansion of the remaining lung or the compensatory increase in extravascular septal tissue volume.     

Increased cardiac output increases shunt: role of pulmonary edema and perfusion

In low-pressure pulmonary edema increased cardiac output (QT) increases shunt (Qs/QT); we tested whether the mechanism is an increase in extravascular lung water in turn mediated by the accompanying increase in microvascular pressure. In six pentobarbital sodium-anesthetized dogs ventilated with O2 we administered oleic acid into the right atrium. From base line to 2 h post-oleic acid we measured concurrent significant increases in Qs/QT (6-29%, O2 technique) and extravascular thermal volume (ETV, 2.6-7.1 ml/g dry intravascular blood-free lung wt, thermal-green dye indicator technique) that were stable by 90 min. Then, bilateral femoral arteriovenous fistulas were opened and closed in 30-min periods to cause reversible increases in QT and associated Qs/QT. When fistulas were open the time-averaged QT increased from 5.1 to 6.9 min (P less than 0.05), the simultaneous Qs/QT rose from 30.7 to 38.4% (P less than 0.05), but ETV did not increase. We conclude that increasing lung edema does not account for our rise in Qs/QT when QT increased.     

A comparative study of postpneumonectomy compensatory lung response in growing male and female rats.

Postpneumonectomy compensatory lung response and normal lung growth in the early postnatal period were studied in male and female rats. Four-week-old litter-matched male and female Sprague-Dawley rats were subjected to left pneumonectomy or sham operation and followed for 3 wk. In both sexes after pneumonectomy, lung weight (WL), lung volume (VL), alveolar surface area (Sw), total alveolar number (N(at)), and the amount of DNA and protein increased significantly. In both males and females, WL, VL, and Sw matched those of both lungs of the sham-operated group, but N(at) and the amount of DNA and protein did not. Female pneumonectomy and sham-operated rats were smaller in body weight than males. Absolute WL, VL, Sw, N(at), and the amount of DNA and protein were significantly lower, but specific parameters (per unit body weight) were significantly greater in females than in males. After pneumonectomy, the postcaval lobe increased most in volume (70 and 73% in males and females, respectively). Mean linear intercept and mean chord length of alveoli increased, and the number of alveoli per unit volume decreased more in the postcaval and middle lobes than in upper and lower lobes in both sexes. Postpneumonectomy, loss of elastic lung recoil was observed in females. We conclude that, in certain aspects (WL, VL), compensatory growth matched both lungs of controls, but in others (biochemical, morphometric) it did not. There was evidence of alveolar multiplication, but the dominant effect was enlargement of air spaces.(ABSTRACT TRUNCATED AT 250 WORDS)     

Lymphokine-activated killer cells with interleukin-2: dose toxicity and localization in isolated perfused rat lungs

Lymphokine-activated killer (LAK) cells combined with recombinant interleukin-2 (rIL-2) can produce tumor regression in murine models and in patients with pulmonary metastatic disease. However, the dose escalations of rIL-2 required for optimal therapeutic effect often result in increased vascular permeability ("vascular leak syndrome") and other toxic systemic consequences. To avoid systemic distribution, lung perfusion was used to administer LAK and rIL-2 locally. Preliminary to using these agents to treat tumor-bearing lungs, we used a nonblood-perfused isolated rat lung model to study the localization of radiolabeled rIL-2 and LAK and to characterize effects on normal lung tissue of increasing dosages and exposure times of rIL-2 and LAK cells, individually and combined. Lung function or permeability was assessed by measuring lung weight gain and pulmonary arterial pressure during the perfusion, extravascular lung water by double indicator dilution techniques, and wet weight to dry weight ratio. After perfusion for 1 hour using 200,000 U (1,300 U/ml) rIL-2, injury was detected as visible pulmonary edema, weight gain and increases in wet to dry weight ratio, and extravascular lung water; no injury was detected at lower, clinically appropriate dosages. When 1 X 10(8) LAK cells combined with 100,000 U rIL-2 (666 U/ml) were perfused for up to 2 hours, no injury was ascertained. Uptake and distribution of the radiolabeled rIL-2 or LAK was uniform to all lung lobes and corresponded to the decrease of 12% of the rIL-2 or 50% of the LAK from the perfusate after 1-hour perfusion.     

Effect of atelectasis and embolization on extravascular thermal volume of the lung

The extravascular volume of distribution for heat in the lung has been advocated for the measurement of lung water. The purpose of these experiments was to investigate how extremes of ventilation-perfusion mismatch influence this measurement. Twenty-six dogs were studied with right and left atrium-to-aorta thermal and dye-dilution curves before and 60 min after total right main-stem bronchial obstruction or microembolization of the pulmonary circulation with 0.275-mm glass beads. Whereas atelectasis had no influence on our measurements, embolization with 0.32 g/kg of beads decreased the detected pulmonary blood volume from 10.63 to 8.55 ml/kg and increased the extravascular thermal volume (ETV) from 9.89 to 10.99 ml/kg. Embolization with 0.65 g/kg decreased the detected ETV from 9.29 to 8.38 ml/kg, while the extravascular wet-to-dry weight ratio was increased, and the regression of postmortem extravascular mass on ETV differed from control. We conclude that microembolization but not atelectasis causes errors in the measurement of lung fluid when the thermodye technique is used. The errors are variable and depend on the degree of embolization.     

Lung mechanics and connective tissue proteins in diabetic Bio-Breeding/Worcester Wistar rats

We studied lungs of spontaneously diabetic Bio-Breeding/Worcester (BB/W) Wistar rats which resemble human insulin-dependent diabetes mellitus. Compared with the age-matched control group, the body weight of the diabetic rats tended to be smaller and lung wet and dry weight were similar, but lung dry weight, relative to body weight and to lung wet weight, was significantly larger. Both air and saline lung volumes were reduced in the diabetic rats, and volume-pressure (V-P) curves expressed as a percent of maximal lung volume were significantly shifted downward and to the right of those in the control group over the midportion. Total DNA and RNA contents were similar in both groups, whereas protein content and concentration and protein/DNA and RNA/DNA ratios were significantly reduced in the diabetic rats. In contrast, content and concentration of 4-hydroxy-L-proline, elastin, and crude connective tissue were significantly higher in the diabetic group. We conclude that the increase in connective tissue proteins in the BB/W rats is most likely responsible for the shift in the V-P curves.     

Pulmonary edema with smoke inhalation, undetected by indicator-dilution technique

Despite experimental evidence for an increase in extravascular lung water (EVLW) after inhalation injury, thermal-dye estimations of EVLW, extravascular thermal volume (EVTV), have repeatedly failed to demonstrate its presence in patients. This situation was evaluated in a sheep model. Under halothane anesthesia one lung was insufflated with cotton smoke and the other with air. EVTV values were 8.4 +/- 0.48 ml/kg at base line and were not elevated at 24 h after smoke inhalation (8.3 +/- 0.45 ml/kg; means +/- SE). Gravimetric analysis 24 h after smoke inhalation showed the development of edema in smoke-exposed lungs. The blood-free wet weight-to-dry weight ratio of the smoke-exposed lungs (5.4 +/- 0.32) was significantly higher compared with the contralateral unsmoked lungs (4.3 +/- 0.15; P less than or equal to 0.05). The thermal-dye technique thus underestimates EVLW. Poor perfusion of the smoke-exposed lungs 24 h after injury was demonstrated indirectly by killing a group of sheep with T-61, an agent that causes a dark red coloration of well-perfused lung areas, as well as directly by measurement of blood flow utilizing a radiolabeled microsphere technique. Thus the inability of the thermal-dye technique to detect the lung edema may be the result of poor perfusion of the injured lung.     

Endothelial injury and pulmonary congestion characterize neurogenic pulmonary edema in rabbits

The objectives of the present study were to determine whether an intracisternal injection of fibrinogen-sodium citrate, a model of neurogenic pulmonary edema (NPE), produces protein-rich or protein-poor pulmonary edema, and to determine whether the edema is associated with pulmonary vascular hypertension and pulmonary congestion. Fibrinogen (6-10 mg/ml) dissolved in 0.055 M sodium citrate was injected into the cisterna magna of six New Zealand White rabbits. Six additional rabbits were injected with saline to control for the effects of intracranial hypertension and pulmonary vascular hypertension. The fibrinogen-sodium citrate solution or sodium citrate alone, as opposed to saline, produced systemic and pulmonary vascular hypertension, pulmonary edema, hypoxemia, hypercapnia, and acidosis. The lungs from fibrinogen-injected rabbits were edematous, congested, and liverlike in appearance. Tracheal froth that was blood tinged and protein rich was present in five of the six rabbits. Microscopic examination of lung biopsies revealed erythrocytes and plasma in the alveoli and focal injury to the pulmonary microvascular endothelium. Fibrinogen-sodium citrate increased (P less than 0.05) the extravascular lung water (EVLW) (10.3 +/- 2.0 vs. 5.5 +/- 0.6 g, means +/- SE), lung blood weight (9.7 +/- 1.3 vs. 3.8 +/- 0.6 g), total dry lung weight (3.2 +/- 0.4 vs. 2.0 +/- 0.1 g), and the EVLW-to-blood-free dry lung weight ratio (7.0 +/- 0.8 vs. 4.0 +/- 0.3 g) from saline-control values. There was no difference in the blood-fre dry lung weight (1.4 +/- 0.1 vs. 1.3 +/- 0.1 g) between the two groups. These findings demonstrate that pulmonary congestion, pulmonary vascular hypertension, and focal endothelial injury contribute to the development of NPE.     

Effects of crystalloid and colloid fluids on extravascular lung water in hypoproteinemic dogs

We compared the effect of crystalloid to colloid fluid infusion on extravascular lung water (EVLW) in hypoproteinemic dogs. Plasmapheresis was used to decrease plasma colloid osmotic pressure (COP) to less than 40% of its base-line level. Five animals were then infused with 0.9% sodium chloride (saline), five with 5% human serum albumin (albumin), and five with 6% hydroxyethyl starch (hetastarch) to increase the pulmonary arterial occlusive pressure by 10 Torr in comparison to the postplasmapheresis level for a 5-h study interval. On completion of the procedure, the lungs were harvested and EVLW measured by the blood-free gravimetric technique. Three to six times the volume of saline compared with albumin or hetastarch (P less than 0.001) was infused. In the saline animals, COP was decreased to 3.3 +/- 1.3 Torr, whereas COP was increased to 18.1 +/- 1.4 Torr in albumin animals (P less than 0.001) and 20.1 +/- 1.6 Torr in the hetastarch group (P less than 0.001). The saline-treated dogs developed gross signs of systemic edema. The EVLW was 8.1 +/- 0.9 ml/kg in saline animals compared with 5.3 +/- 2.1 ml/kg in the albumin (P less than 0.05) and 4.1 +/- 1.4 ml/kg in the hetastarch (P less than 0.01) groups. These data indicate that crystalloid fluid infusion during hypoproteinemia is associated with the development of both systemic and pulmonary edema.     

Effect of bronchial artery blood flow on cardiopulmonary bypass-induced lung injury

Cardiovascular surgery requiring cardiopulmonary bypass (CPB) is frequently complicated by postoperative lung injury. Bronchial artery (BA) blood flow has been hypothesized to attenuate this injury. The purpose of the present study was to determine the effect of BA blood flow on CPB-induced lung injury in anesthetized pigs. In eight pigs (BA ligated) the BA was ligated, whereas in six pigs (BA patent) the BA was identified but left intact. Warm (37 degrees C) CPB was then performed in all pigs with complete occlusion of the pulmonary artery and deflated lungs to maximize lung injury. BA ligation significantly exacerbated nearly all aspects of pulmonary function beginning at 5 min post-CPB. At 25 min, BA-ligated pigs had a lower arterial Po(2) at a fraction of inspired oxygen of 1.0 (52 +/- 5 vs. 312 +/- 58 mmHg) and greater peak tracheal pressure (39 +/- 6 vs. 15 +/- 4 mmHg), pulmonary vascular resistance (11 +/- 1 vs. 6 +/- 1 mmHg x l(-1) x min), plasma TNF-alpha (1.2 +/- 0.60 vs. 0.59 +/- 0.092 ng/ml), extravascular lung water (11.7 +/- 1.2 vs. 7.7 +/- 0.5 ml/g blood-free dry weight), and pulmonary vascular protein permeability, as assessed by a decreased reflection coefficient for albumin (sigma(alb); 0.53 +/- 0.1 vs. 0.82 +/- 0.05). There was a negative correlation (R = 0.95, P < 0.001) between sigma(alb) and the 25-min plasma TNF-alpha concentration. These results suggest that a severe decrease in BA blood flow during and after warm CPB causes increased pulmonary vascular permeability, edema formation, cytokine production, and severe arterial hypoxemia secondary to intrapulmonary shunt.     

Endotoxin increases pulmonary vascular protein permeability in the dog

Endotoxin increases pulmonary vascular permeability consistently in some species but fails to reliably cause injury in the dog. We wondered whether this phenomenon depended on the method of injury assessment, as others have relied on edema measurement; we quantified injury by monitoring the rate of extravascular protein accumulation. 113mIn-labeled protein and 99mTc-labeled erythrocytes were injected into anesthetized dogs and monitored by an externally placed lung probe. A protein leak index, the rate of extravascular protein accumulation, was derived from the rate of increase in lung protein counts corrected for changes in intravascular protein activity. After administration of Salmonella enteriditis endotoxin (4 micrograms/kg), the protein leak index was elevated 2.5-fold (41.1 +/- 4.6 X 10(-4) min-1) compared with control (16.0 +/- 2.8 X 10(-4) min-1). In contrast, wet-to-dry weight ratios failed to increase after endotoxin (4.6 +/- 0.8 vs. control values of 4.2 +/- 0.5 g/g dry bloodless lung). However, we observed that endotoxin increased lung dry weight (per unit body weight), which may have attenuated the change in wet-to-dry weight ratios. To determine whether low microvascular pressures following endotoxin attenuated edema formation, we increased pulmonary arterial wedge pressures in five dogs by saline infusion, which caused an increase in wet-to-dry weight ratios following endotoxin but no change in the five controls. We conclude that low dose endotoxin causes pulmonary vascular protein leak in the dog while edema formation is minimal or absent.     

Spontaneous injury in isolated sheep lungs: role of perfusate leukocytes and platelets

Perfusion of isolated sheep lungs with blood causes spontaneous edema and hypertension preceded by decreases in perfusate concentrations of leukocytes (WBC) and platelets (PLT). To determine whether these decreases were caused by pulmonary sequestration, we continuously measured blood flow and collected pulmonary arterial and left atrial blood for cell concentration measurements in six lungs early in perfusion. Significant sequestration occurred in the lung, but not in the extracorporeal circuit. To determine the contribution of these cells to spontaneous injury in this model, lungs perfused in situ with a constant flow (100 ml.kg-1.min-1) of homologous leukopenic (WBC = 540 mm-3, n = 8) or thrombocytopenic blood (PLT = 10,000 mm-3, n = 6) were compared with control lungs perfused with untreated homologous blood (WBC = 5,320, PLT = 422,000, n = 8). Perfusion of control lungs caused a rapid fall in WBC and PLT followed by transient increases in pulmonary arterial pressure, lung lymph flow, and perfusate concentrations of 6-ketoprostaglandin F1 alpha and thromboxane B2. The negative value of reservoir weight (delta W) was measured as an index of fluid entry into the lung extravascular space during perfusion. delta W increased rapidly for 60 min and then more gradually to 242 g at 180 min. This was accompanied by a rise in the lymph-to-plasma oncotic pressure ratio (pi L/pi P). Relative to control, leukopenic perfusion decreased the ratio of wet weight to dry weight, the intra- plus extravascular blood weight, and the incidence of bloody lymph. Thrombocytopenic perfusion increased lung lymph flow and the rate of delta W, decreased pi L/pi P and perfusate thromboxane B2, and delayed the peak pulmonary arterial pressure. These results suggest that perfusate leukocytes sequestered in the lung and contributed to hemorrhage but were not necessary for hypertension and edema. Platelets were an important source of thromboxane but protected against edema by an unknown mechanism.