Electron microscopy of lung rapidly frozen under controlled physiological conditions

Light microscopy of lung rapidly frozen under controlled physiological conditions has been very successful in correlating pulmonary structure and function. However, to study some aspects of pulmonary capillary morphology, the higher resolution of electron microscopy (EM) is necessary. To date, most EM of lung has involed the instillation of a fixative through the airways or vascular system, techniques that probably alter the normal pressure relationships of the capillaries and therefore their morphology. We describe here a technique for rapidly freezing lung to a depth of 1--2 mm below the pleural surface and preparing sections for EM. Lungs from open-chest rats were frozen at various transpulmonary pressures with cold (--80 degrees C) 70% ethylene glycol. Small pieces were then fixed with a solution containing glutaraldehyde and paraformaldehyde for 24 h at --50 degrees C. Staining was with osmium tetroxide and uranyl acetate. Lung frozen at high volumes showed marked stretching of the alveolar septa with severe deformation of the capillaries. Lung frozen at low inflation pressures revealed open capillaries containing numerous red blood cells; in addition, infolding of the alveolar wall was frequently seen. We conclude that this technique gives a level of preservation of rapidly frozen lung suitable for electron microscopy.     

Red blood cell orientation in pulmonary capillaries and its effect on gas diffusion.

When alveoli are inflated, the stretched alveolar walls draw their capillaries into oval cross sections. This causes the disk-shaped red blood cells to be oriented near alveolar gas, thereby minimizing diffusion distance. We tested these ideas by measuring red blood cell orientation in histological slides from rapidly frozen rat lungs. High lung inflation did cause the capillaries to have oval cross sections, which constrained the red blood cells within them to flow with their broad sides facing alveolar gas. Low lung inflation stretched alveolar walls less and allowed the capillaries to assume a circular cross section. The circular luminal profile permitted the red blood cells to have their edges facing alveolar gas, which increased the diffusion distance. Using a finite-element method to calculate the diffusing capacity of red blood cells in the broad-side and edge-on orientations, we found that edge-on red blood cells had a 40% lower diffusing capacity. This suggests that, when capillary cross sections become circular, whether through low-alveolar volume or through increased microvascular pressure, the red blood cells are likely to be less favorably oriented for gas exchange.     

The influence of maternal nicotine exposure on the interalveolar septal status of neonatal rat lung.

The aim of this study was to determine the effect of maternal nicotine exposure (1 mg nicotine/kg body mass/day, subcutaneously) on the status of the alveolar septa of the 1 to 21 day old offspring. The data obtained showed swelling of type II and interstitial cell mitochondria. The type I:type II cell ratio decreased as a result of type II cell proliferation. The number of capillaries per unit length of septum was also significantly lower than that of control lung. Ruptured blood-air barriers also occur in the nicotine exposed lungs of rats of all age groups. The results show that maternal nicotine exposure interfered with the morphometric and morphologic characteristics of the septa of lung tissue of the offspring.     

Status of the capillary endothelium of the lung in normovolemic perfusion of the organ through the system of the pulmonary artery

Pulmonary perfusion for 30 min to the dog under conditions of normovolemia is not accompanied with any essential changes in parameters of alveolar capillaries endothelium. Just the opposite, transformation of endothelial lining of the peribronchial capillaries demonstrates possible disturbances of the liquor transport across the walls of these vessels. The volumetric part of the interstitial space near these capillaries increases, while in the alveolar septa it does not change. In lymph formation, flowing out of the lung, together with bronchial capillaries, blood capillaries of the alveoli must take part.     

Diaspirin cross-linked hemoglobin improves oxygen extraction capabilities in endotoxic shock.

We studied the effects of diaspirin cross-linked hemoglobin (DCLHb), a cell-free hemoglobin derived from human erythrocytes, on blood flow distribution and tissue oxygen extraction capabilities in endotoxic shock. Eighteen pentobarbital sodium-anesthetized, mechanically ventilated dogs received 2 mg/kg of E. coli endotoxin, followed by saline resuscitation to restore cardiac filling pressures to baseline levels. The animals were randomly divided into three groups: six served as control, six received DCLHb at a dose of 500 mg/kg (group 1) and six DCLHb at a dose of 1,000 mg/kg (group 2). Cardiac tamponade was then induced by saline injection in the pericardial sac to progressively reduce cardiac index and thereby allow study of tissue oxygen extraction capabilities. DCLHb had a dose-dependent vasopressor effect but did not significantly alter cardiac index or regional blood flow. During cardiac tamponade, critical oxygen delivery was 12.8 +/- 0.7 ml. kg(-1). min(-1) in the control group, but 8.6 +/- 0.9 and 8.2 +/- 0.7 ml. kg(-1). min(-1) in groups 1 and 2, respectively (both P < 0.05 vs. control group). The critical oxygen extraction ratio was 39.1 +/- 3.1% in the control group but 58.7 +/- 12.8% and 60.2 +/- 9.0% in groups 1 and 2, respectively. We conclude that DCLHb can improve whole body oxygen extraction capabilities during endotoxic shock in dogs.     

Short-term reversibility of ultrastructural changes in pulmonary capillaries caused by stress failure.

We previously showed that when the pulmonary capillaries in anesthetized rabbits are exposed to a transmural pressure (Ptm) of approximately 40 mmHg, stress failure of the walls occurs with disruption of the capillary endothelium, alveolar epithelium, or sometimes all layers. The present study was designed to determine whether some of the ultrastructural changes are rapidly reversible when the capillary pressure is reduced. To test this, the Ptm was raised to 52.5 cmH2O for 1 min of blood perfusion and then reduced to 12.5 cmH2O for 3 min of saline-dextran perfusion, followed by intravascular fixation at the same pressure. In another group of animals, the pressure was elevated for 1 min of blood and 3 min of saline-dextran before being reduced. The results were compared with previous studies in which the capillary pressures were maintained elevated at 52.5 cmH2O during the entire procedure. Control studies were also done at sustained low pressures. The results showed that the number of endothelial and epithelial breaks per millimeter and the total fraction area of the breaks were reduced when the pressure was lowered. For example, the number of endothelial breaks per millimeter decreased from 7.1 +/- 2.1 to 2.4 +/- 0.7, and the number of epithelial breaks per millimeter fell from 11.4 +/- 3.7 to 3.4 +/- 0.7. There was evidence that the breaks that closed were those that were initially small and were associated with an intact basement membrane. The results suggest that cells can move along their underlying matrix by rapid disengagement and reattachment of cell adhesion molecules, causing breaks to open or close within minutes.(ABSTRACT TRUNCATED AT 250 WORDS)     

A scanning and transmission electron microscopic study of the bat lung

The lungs of two adult species of bat Epomophorus wahlbergi and Miniopterus minor fixed with 2.3% glutaraldehyde were processed for SEM (scanning electron microscope) and TEM (transmission electron microscope) examination by the standard procedures. The bat lung comprised a blood and air conducting zone (consisting of bronchi, bronchioles and large blood vessels), the intermediate zone (made up of alveolar ducts), and the respiratory zone, which consisted of alveoli and blood capillaries. The interalveolar septa comprised basically granular pneumocytes (type II cells), squamous pneumocytes (type I cells), endothelial cells, and, in the interstitium, collagen and elastic fibres with occasional fibrocytes. Blood capillaries were interposed in the interalveolar septa, thus bulging into adjacent alveoli. It was noted that grossly, architecturally and structurally, the bat lung was similar to that of a terrestrial mammal. However, in previous morphometric and physiological studies it has been found that bats have a large lung, a thin pulmonary blood-gas barrier, a large pulmonary capillary blood volume, and high haematocrit and haemoglobin concentration. The bat lung, while retaining the basic mammalian pulmonary design, is well adapted to provide the large amount of oxygen demanded by flight. The avian pulmonary design (the lung-air sac system) is thus not a prerequisite to flight.     

Effect of alveolar and pleural pressures on interstitial pressures in isolated dog lungs

We report the first direct measurements of perialveolar interstitial pressures in lungs inflated with negative pleural pressure. In eight experiments, we varied surrounding (pleural) pressure in a dog lung lobe to maintain constant inflation with either positive alveolar and ambient atmospheric pleural pressures (positive inflation) or ambient atmospheric alveolar and negative pleural pressures (negative inflation). Throughout, vascular pressure was approximately 4 cmH2O above pleural pressure. By the micropuncture servo-null technique we recorded interstitial pressures at alveolar junctions (Pjct) and in the perimicrovascular adventitia (Padv). At transpulmonary pressure of 7 cmH2O (n = 4), the difference of Pjct and Pady from pleural pressure of 0.9 +/- 0.4 and -1.1 +/- 0.2 cmH2O, respectively, during positive inflation did not significantly change (P less than 0.05) after negative inflation. After increase of transpulmonary pressure from 7 to 15 cmH2O (n = 4), the decrease of Pjct by 3.3 +/- 0.3 cmH2O and Pady by 2.0 +/- 0.4 cmH2O during positive inflation did not change during negative inflation. The Pjct-Pady gradient was not affected by the mode of inflation. Our measurements indicate that, in lung, when all pressures are referred to pleural or alveolar pressure, the mode of inflation does not affect perialveolar interstitial pressures.     

High lung volume increases stress failure in pulmonary capillaries.

We previously showed that when pulmonary capillaries in anesthetized rabbits are exposed to a transmural pressure (Ptm) of approximately 40 mmHg, stress failure of the walls occurs with disruption of the capillary endothelium, alveolar epithelium, or sometimes all layers. The present study was designed to test whether stress failure occurred more frequently at high than at low lung volumes for the same Ptm. Lungs of anesthetized rabbits were inflated to a transpulmonary pressure of 20 cmH2O, perfused with autologous blood at 32.5 or 2.5 cmH2O Ptm, and fixed by intravascular perfusion. Samples were examined by both transmission and scanning electron microscopy. The results were compared with those of a previous study in which the lung was inflated to a transpulmonary pressure of 5 cmH2O. There was a large increase in the frequency of stress failure of the capillary walls at the higher lung volume. For example, at 32.5 cmH2O Ptm, the number of endothelial breaks per millimeter cell lining was 7.1 +/- 2.2 at the high lung volume compared with 0.7 +/- 0.4 at the low lung volume. The corresponding values for epithelium were 8.5 +/- 1.6 and 0.9 +/- 0.6. Both differences were significant (P less than 0.05). At 52.5 cmH2O Ptm, the results for endothelium were 20.7 +/- 7.6 (high volume) and 7.1 +/- 2.1 (low volume), and the corresponding results for epithelium were 32.8 +/- 11.9 and 11.4 +/- 3.7. At 32.5 cmH2O Ptm, the thickness of the blood-gas barrier was greater at the higher lung volume, consistent with the development of more interstitial edema. Ballooning of the epithelium caused by accumulation of edema fluid between the epithelial cell and its basement membrane was seen at 32.5 and 52.5 cmH2O Ptm. At high lung volume, the breaks tended to be narrower and fewer were oriented perpendicular to the axis of the pulmonary capillaries than at low lung volumes. Transmission and scanning electron microscopy measurements agreed well. Our findings provide a physiological mechanism for other studies showing increased capillary permeability at high states of lung inflation.     

Effect of dehydration on interstitial pressures in the isolated dog lung

We have determined the effect of dehydration on regional lung interstitial pressures. We stopped blood flow in the isolated blood-perfused lobe of dog lung at vascular pressure of approximately 4 cmH2O. Then we recorded interstitial pressures by micropuncture at alveolar junctions (Pjct), in perimicrovascular adventitia (Padv), and at the hilum (Phil). After base-line measurements, we ventilated the lobes with dry gas to decrease extravascular lung water content by 14 +/- 5%. In one group (n = 10), at constant inflation pressure of 7 cmH2O, Pjct was 0.2 +/- 0.8 and Padv was -1.5 +/- 0.6 cmH2O. After dehydration the pressures fell to -5.0 +/- 1.0 and -5.3 +/- 1.3 cmH2O, respectively (P less than 0.01), and the junction-to-advential gradient (Pjct-Padv) was abolished. In a second group (n = 6) a combination of dehydration and lung expansion with inflation pressure of 15 cmH2O further decreased Pjct and Padv to -7.3 +/- 0.7 and -7.1 +/- 0.7 cmH2O, respectively. Phil followed changes in Padv. Interstitial compliance was 0.6 at the junctions, 0.8 in adventitia, and 0.9 ml.cmH2O-1.100 g-1 wet lung at the hilum. We conclude, that perialveolar interstitial pressures may provide an important mechanism for prevention of lung dehydration.     

Longitudinal distribution of vascular compliance in the canine lung

With an isolated perfused canine lung, the compliance of pulmonary circulation was measured and partitioned into components corresponding to alveolar and extra-alveolar compartments. When the lungs were in zone 3, changes in outflow pressure (delta Po) affected all portions of the vasculature causing a change in lung blood volume (delta V). Thus the ratio delta V/delta Po in zone 3 represented the compliance of the entire pulmonary circulation (Cp) plus that of the left atrium (Cla). When the lungs were in zone 2, changes in Po affected only the extra-alveolar vessels that were downstream from the site of critical closure in the alveolar vessels. Thus the ratio delta V/delta Po with forward flow in zone 2 represented the compliance of the venous extra-alveolar vessels (Cv) plus Cla. With reverse flow in zone 2, delta V/delta Po represented the compliance of the arterial extra-alveolar vessels (Ca). The compliance of the alveolar compartment (Calv) was calculated from the difference between Cp and the sum of Ca + Cv. When Po was 6-11 mmHg, Cp was 0.393 +/- 0.0380 (SE) ml X mmHg-1 X kg-1 with forward perfusion and 0.263 +/- 0.0206 (SE) ml X mmHg-1 X kg-1 with reverse perfusion. Calv was 79 and 68% of Cp with forward and reverse perfusion, respectively. When Po was raised to 16-21 mmHg, Cp decreased to 0.225 +/- 0.0235 (SE) ml X mmHg-1 X kg-1 and 0.183 +/- 0.0133 (SE) ml X mmHg-1 X kg-1 with forward and reverse perfusion, respectively. Calv also decreased but remained the largest contributor to Cp. We conclude that the major site of pulmonary vascular compliance in the canine lung is the alveolar compartment, with minor contributions from the arterial and venous extra-alveolar segments.     

Effects of lung volume and alveolar surface tension on pulmonary vascular resistance

Utilizing the arterial and venous occlusion technique, the effects of lung inflation and deflation on the resistance of alveolar and extraalveolar vessels were measured in the dog in an isolated left lower lobe preparation. The lobe was inflated and deflated slowly (45 s) at constant speed. Two volumes at equal alveolar pressure (Palv = 9.9 +/- 0.6 mmHg) and two pressures (13.8 +/- 0.8 mmHg, inflation; 4.8 +/- 0.5 mmHg, deflation) at equal volumes during inflation and deflation were studied. The total vascular pressure drop was divided into three segments: arterial (delta Pa), middle (delta Pm), and venous (delta Pv). During inflation and deflation the changes in pulmonary arterial pressure were primarily due to changes in the resistance of the alveolar vessels. At equal Palv (9.9 mmHg), delta Pm was 10.3 +/- 1.2 mmHg during deflation compared with 6.8 +/- 1.1 mmHg during inflation. At equal lung volume, delta Pm was 10.2 +/- 1.5 mmHg during inflation (Palv = 13.8 mmHg) and 5.0 +/- 0.7 mmHg during deflation (Palv = 4.8 mmHg). These measurements suggest that the alveolar pressure was transmitted more effectively to the alveolar vessels during deflation due to a lower alveolar surface tension. It was estimated that at midlung volume, the perimicrovascular pressure was 3.5-3.8 mmHg greater during deflation than during inflation.     

Blood flow switching among pulmonary capillaries is decreased during high hematocrit.

Pulmonary capillary perfusion within a single alveolar wall continually switches among segments, even when large-vessel hemodynamics are constant. The mechanism is unknown. We hypothesize that the continually varying size of plasma gaps between individual red blood cells affects the likelihood of capillary segment closure and the probability of cells changing directions at the next capillary junction. We assumed that an increase in hematocrit would decrease the average distance between red blood cells, thereby decreasing the switching at each capillary junction. To test this idea, we observed 26 individual alveolar capillary networks by using videomicroscopy of excised canine lung lobes that were perfused first at normal hematocrit (31-43%) and then at increased hematocrit (51-62%). The number of switches decreased by 38% during increased hematocrit (P < 0.01). These results support the idea that a substantial part of flow switching among pulmonary capillaries is caused by the particulate nature of blood passing through a complex network of tubes with continuously varying hematocrit.     

Impact of microvascular circulation on peripheral lung stability

The involvement of pulmonary circulation in the mechanical properties was studied in isolated rat lungs. Pulmonary input impedance (ZL) was measured at a mean transpulmonary pressure (Ptpmean) of 2 cmH2O before and after physiological perfusion with either blood or albumin. In these lungs and in a group of unperfused lungs, ZL was also measured at Ptpmean values between 1 and 8 cmH2O. Airway resistance (Raw) and parenchymal damping (G) and elastance (H) were estimated from ZL. End-expiratory lung volume (EELV) was measured by immersion before and after blood perfusion. The orientation of the elastin fibers relative to the basal membrane was assessed in additional unperfused and blood-perfused lungs. Pressurization of the pulmonary capillaries significantly decreased H by 31.5 +/- 3.7% and 18.7 +/- 2.7% for blood and albumin, respectively. Perfusion had no effect on Raw but markedly altered the Ptpmean dependences of G and H < 4 cmH2O, with significantly lower values than in the unperfused lungs. At a Ptpmean of 2 cmH2O, EELV increased by 31 +/- 11% (P = 0.01) following pressurization of the capillaries, and the elastin fibers became more parallel to the basal membrane. Because the organization of elastin fibers results in smaller H values of the individual alveolus, the higher H in the unperfused lungs is probably due to a partial alveolar collapse leading to a loss in lung volume. We conclude that the physiological pressure in the pulmonary capillaries is an important mechanical factor in the maintenance of the stability of the alveolar architecture.     

Alveolar liquid pressures in newborn and adult rabbit lungs

To study the effects of lung maturation and inflation on alveolar liquid pressures, we isolated lungs from adult and newborn rabbit pups (1-11 days old). We used the micropuncture technique to measure alveolar liquid pressure at several transpulmonary pressures on lung deflation. Alveolar liquid pressure was greater than pleural pressure but less than airway pressure at all transpulmonary pressures. Alveolar liquid pressure decreased further below airway pressure with lung inflation. At high transpulmonary pressure, alveolar liquid pressure was less in newborn than in adult lungs. To study the effects of edema, we measured alveolar liquid pressures in newborn lungs with different wet-to-dry weight ratios. Alveolar liquid pressure increased with progressive edema. In addition, we compared alveolar liquid and perivenular interstitial pressures in perfused newborn lungs and found that they were similar. Thus alveolar liquid pressure can be used to estimate perivenular interstitial pressure. We conclude that the transvascular pressure gradient for fluid flux into the interstitium might increase with lung inflation and decrease with progressive edema. Furthermore, this gradient might be greater in newborn than adult lungs at high inflation pressures.     

Effects of primary resuscitation from shock on distribution of myocardial blood flow.

Hemorrhagic shock alters heterogeneity of regional myocardial perfusion (RMP) in the presence of critical coronary stenosis in pigs. Conventional resuscitation has failed to reverse these effects. We hypothesized that improvement of the resuscitation regime would lead to restoration of RMP heterogeneity. Diaspirin-cross-linked hemoglobin (10 g/dl; DCLHb) and human serum albumin (8.0 g/dl; HSA) were used. After baseline, a branch of the left coronary artery was stenosed; thereafter, hemorrhagic shock was induced. Resuscitation was performed with either DCLHb or HSA. At baseline, the fractcal dimension (D) of subendocardial myocardium was 1.31 +/- 0.083 (HSA) and 1.35 +/- 0.106 (DCLHb) (mean +/- SD). Coronary stenosis increased subendocardial D slightly but consistently only in the DCLHb group (1.39 +/- 0.104; P < 0.05). Shock reduced subendocardial D: 1.21 +/- 0.093 (HSA; P = 0.10), 1.25 +/- 0.092 (DCLHb; P < 0.05). Administration of DCLHb increased subendocardial D in 7 of 10 animals (1.31 +/- 0.097; P = 0.066). HSA was ineffective in this respect. DCLHb infusion restored arterial pressure and increased cardiac index (CI) to 80% of baseline values. Administration of HSA left animals hypotensive (69 mmHg) and increased CI to 122% of the average baseline value. Shock-induced disturbances of the distribution of RMP were improved by administration of DCLHb but not by HSA.     

Seasonal variation in various blood indices of the kangaroo rat, Dipodomys panamintinus

1. Various blood indices in the Panamint kangaroo rat revealed seasonal fluctuations. The red blood cell count during winter and summer averaged 7.2 +/- 1.0 X 10(6) and 9.2 +/- 0.2 X 10(6)/mm3 respectively. 2. The mean cell hemoglobin during winter and summer averaged 25 +/- 10.8 pg and 18.6 +/- 3.7 pg respectively. 3. These fluctuations may reveal a rapid rate of red blood cell destruction during winter in combination with a change in diet, concomitant to this, is an increase in mean cell hemoglobin of the surviving red blood cells.     

Atomic force microscope elastography reveals phenotypic differences in alveolar cell stiffness.

To understand the connection between alveolar mechanics and key biochemical events such as surfactant secretion, one first needs to characterize the underlying mechanical properties of the lung parenchyma and its cellular constituents. In this study, the mechanics of three major cell types from the neonatal rat lung were studied; primary alveolar type I (AT1) and type II (AT2) epithelial cells and lung fibroblasts were isolated using enzymatic digestion. Atomic force microscopy indentation was used to map the three-dimensional distribution of apparent depth-dependent pointwise elastic modulus. Histograms of apparent modulus data from all three cell types indicated non-Gaussian distributions that were highly skewed and appeared multimodal for AT2 cells and fibroblasts. Nuclear stiffness in all three cell types was similar (2.5+/-1.0 kPa in AT1 vs. 3.1+/-1.5 kPa in AT2 vs. 3.3+/-0.8 kPa in fibroblasts; n=10 each), whereas cytoplasmic moduli were significantly higher in fibroblasts and AT2 cells (6.0+/-2.3 and 4.7+/-2.9 kPa vs. 2.5+/-1.2 kPa). In both epithelial cell types, actin was arranged in sparse clusters, whereas prominent actin stress fibers were observed in lung fibroblasts. No systematic difference in actin or microtubule organization was noted between AT1 and AT2 cells. Atomic force microscope elastography, combined with live-cell fluorescence imaging, revealed that the stiffer measurements in AT2 cells often colocalized with lamellar bodies. These findings partially explain reported heterogeneity of alveolar cell deformation during in situ lung inflation and provide needed data for better understanding of how mechanical stretch influences surfactant release.     

Comparison of transfusion with DCLHb or pRBCs for treatment of intraoperative anemia in sheep.

Isoflurane-anesthetized sheep were transfused with packed red blood cells (pRBCs) or diaspirin cross-linked hemoglobin (DCLHb) for treatment of intraoperative hemorrhage. A rapid 15-min hemorrhage with lactated Ringer (LR) infusion maintained filling pressure at baseline and reduced blood hemoglobin (Hb) to ~5 g/dl. Sheep received 2 g/kg Hb, DCLHb (n = 6), or pRBCs (n = 7); control group received LR alone (n = 6). After 2 h, anesthesia was discontinued; sheep were monitored in the animal intensive care unit for 48 h. DCLHb expanded blood volume more, but increased total blood Hb less, than pRBCs. Lower Hb and increased methemoglobin resulted in lower arterial oxygen content compared with the pRBCs. DCLHb caused pulmonary hypertension (from 13 to 30 mmHg) and elevated filling pressure (from 6 to 15 mmHg). Cardiac outputs (CO) were similar for all groups during anesthesia; however, during recovery CO increased only in the LR and packed pRBCs groups. DCLHb may limit the reflex ability to increase CO after volume expansion. Hemodynamic effects of DCLHb may be exaggerated when infused after large-volume LR.     

Capillary perfusion patterns in single alveolar walls.

We studied capillary perfusion patterns in single alveolar walls through a transparent thoracic window implanted in pentobarbital-anesthetized dogs. The capillaries were maximally opened by brief inflation of a balloon in the left atrium to raise pressure. After the balloon was deflated and pulmonary hemodynamics returned to zone 2 baseline conditions, the capillaries that remained perfused in the observed field were videotaped with the use of in vivo microscopy. The cycle of elevated pressure and baseline observation was repeated three times. Perfusion of different capillaries during each of the observations would imply that the capillaries had characteristics that permitted flow to switch between segments. Perfusion of a specific set of pathways through the network each time would demonstrate that flowing blood sought a unique and repeatable combination of segments, presumably with the least total pathway resistance. We found that the same capillary segments were perfused 79% of the time, a strong indication that a reproducible combination of individual segmental resistances determined the predominant pattern of pulmonary capillary perfusion.