INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

GRAVITY INDUCES A HYDROSTATIC gradient down the lung that distends dependent vessels and influences regional blood-flow distribution (25). The pulmonary vascular system, however, is not infinitely compliant, and the geometry of the vascular tree must also play a role in perfusion distribution. To determine the relative contributions of gravity and vascular structure to the distribution of pulmonary blood flow, we flew with pigs on the National Aeronautics and Space Administration (NASA) KC-135 aircraft. The flight path was a series of 40 parabolas that created alternating weightlessness and 1.8-G conditions. After a steady state was reached in each gravitational condition, fluorescent microspheres of varying colors were injected into the pulmonary circulation to mark regional blood flow. The lungs were subsequently removed, air dried, and sectioned into small pieces. By analyzing the dye content in each piece, we obtained high-resolution measurements of pulmonary blood-flow distribution during 0-, 1-, and 1.8-G conditions in supine and prone postures.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The Animal Care Committees at the University of Washington, NASA, and Baylor University approved these animal studies. Six juvenile pigs weighing between 23 and 25 kg were studied. They were purpose bred and pathogen free (S and S Farms, Ranchita, CA). All animals were cared for and handled in accordance with the guidelines established by the National Institutes of Health (17).

Experiments were performed on the NASA KC-135 microgravity research aircraft. A series of concave and convex parabolas were flown to provide alternating weightless and increased gravitational conditions (Fig. 1). The trajectory of the aircraft produced three consecutive gravitational phases of 0-, ~1.5-, and 1.8-G conditions. The duration of weightlessness and 1.8-G conditions varied slightly with each parabola but were at least 23 s long. A series of 10 parabolic loops were flown with 3-5 min of level flight between each series. A total of 40 parabolas were flown during each flight and were completed over ~1 h. Normal gravitational conditions existed before and in between the 10-loop parabolic cycles. The cabin was pressurized to an average of 626 Torr but varied with changes in altitude.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Flight trajectory and gravitational forces during parabolic flight on the NASA KC-135. Repeating parabolic loops create alternating weightless and 1.8-G conditions lasting ~25 s each. A series of 10 parabolic loops were flown, with 3-5 min of level flight between each series; 40 parabolas were flown for each flight (based on similar schematic shown in Ref. 2a).

Physiological monitoring and data recording. Physiological and environmental variables were monitored and recorded continuously with the use of a portable PowerPC Macintosh (Apple Computer) with a MacLab/8s (ADInstruments, Castle Hill, NSW, Australia). All analog signals were sampled at 40 signals/s. Airway (Paw) and esophageal pressures (Pes) were measured with Valdyne CD-15 transducers (Northridge, CA). Intravascular pressures were measured with Abbott Critical Care Systems transducers (Chicago, IL). Electrocardiogram, end-tidal CO₂ concentrations, and arterial O₂ saturation were measured with the use of a Pace-tech 3100 (Clearwater, FL). The distance from the midesophagus to the back of the animal was continuously measured with a linear Hall-effect sensor (Allegro, A3515EUA). The Hall-effect sensor was located at the tip of the esophageal catheter and generated a signal proportional to the distance between it and a magnetic source fixed on the back of the animal. The sensor and magnet were kindly provided by Lucent Medical Systems (Seattle, WA). The magnitude of the gravitational force was continuously measured with a calibrated strain-gauge force transducer (FT03C, Grass Instruments, West Warwick, RI). An electronic signal was generated at the beginning and end of important events to correlate later with gravitational conditions. Cardiac outputs were measured by thermal dilution using a Baxter SAT-2 cardiac output monitor (Baxter-Edwards Health Care, Irvine, CA). The direction of the gravitational vector relative to the floor of the aircraft was measured with the use of a digital level. The cabin pressure was measured with an analog altimeter. These last three measurements were recorded by hand at appropriate times. Arterial blood-gas samples were collected in glass syringes during scheduled conditions, placed on ice, and analyzed with an IRMA blood-analysis system (Diametrics Medical, St. Paul, MN) after we returned to base.

Fluorescent microspheres. Eleven different colors of polystyrene microspheres were obtained from Molecular Probes (Eugene, OR) and Bangs Laboratory (Indianapolis, IN). The colors used were blue, blue-green, green, yellow-green, yellow, orange, orange-red, red, crimson, scarlet, and carmine. The microspheres had diameters of 15.5 µm and a specific gravity of 1.05. The microspheres were suspended in a 17.3% dextrose solution with a specific gravity of 1.05 so that they would not settle during parabolic flights. An antibacterial agent, Thimerisol, was added to the suspension. Just before each flight, the microspheres were sonicated, vortexed, and drawn into labeled syringes. Each syringe contained one million microspheres of a single color in a 1.25-ml volume. The order in which the colors were injected was random and different for each flight except for the last two injections, which were always blue or carmine (see below).

Preflight. The animals were not fed for 12 h before experimentation but had free access to water. Ninety minutes before flight time, each pig was sedated with an intramuscular injection of ketamine (20 mg/kg) and xylazine (2 mg/kg). The animals were anesthetized with a continuous intravenous infusion of thiopental, and the infusion rate was adjusted to suppress spontaneous respirations and maintain a surgical plane of anesthesia. A tracheotomy was performed, and the animals were ventilated with 30% O₂ using an Auto-vent 3000 (Life Support Products, St. Louis, MO). Tidal volumes of 210-240 ml were used, and the rate was set to hyperventilate the animals with a targeted arterial PCO₂ of ~33 Torr. Paw were measured from a side port of the tracheal tube. O₂ saturation was measured with an optical sensor clipped to the ear of the animal. Internal carotid, internal jugular, and pulmonary arterial catheters were placed, and their positions were verified by pressure tracings. The internal jugular catheter had three separate ports and lumens. The most distal lumen, used for microsphere injections, was placed just proximal to the right atrium. An esophageal balloon catheter with a Hill-effect transistor at the tip was passed so that the tip was in a retrocardiac position and the balloon was partially inflated. A magnet was taped over the region of the thoracic spine that produced the largest signal from the esophageal Hill-effect transistor. This signal provided an estimate of the mediastinal shift in the sagittal plane during changes in the gravitational conditions.

In-flight. Paw, Pes, central venous pressure, pulmonary artery pressure, and systemic artery pressure were continuously monitored and recorded throughout the flight (Fig. 2). The animals were ventilated with 30% O₂ and given stacked breaths after each 1.8-G condition to minimize atelectasis. Anesthesia was adjusted as necessary.

View larger version (48K): [in this window] [in a new window]   Fig. 2.   Sample of the digitized signals recorded continuously throughout the flight. Example shows a full parabolic loop, including weightlessness, a transition phase, and 1.8 G over 100 s. Airway (Paw), esophageal (Pes), central venous (CVP), pulmonary arterial (Ppa), and systemic arterial (Psa) pressures are shown. Temporal course of G conditions (G-force) is recorded at bottom. Marker tracing indicates the beginning (A) and end (B) of a microsphere injection and time when cardiac output was measured (C).

Postflight. The arterial blood-gas samples were analyzed shortly after we landed. The animals were deeply anesthetized, exsanguinated, and given an overdose of intravenous thiopental. A sternotomy was performed, large-bore catheters were placed in the pulmonary artery and left atrium, and the thoracic aorta was tied off. The lungs were perfused with 2% Dextran (molecular weight of 74,000) in normal saline until the effluent was clear of blood, removed from the chest, and allowed to dry inflated at a Paw of 25 cmH₂O.

Statistics. All data are presented as means ± SD except where noted. The goodness of the linear fit, r², between two variables was used to quantify the strength of the relationship. We used ANOVA for statistical comparisons with a randomized complete block design of posture and gravitational force. This design allowed us to evaluate the main effects of posture and gravitational conditions and the interaction effect between them. Statistical significance was declared for P < 0.05. The coefficient of variation (CV = SD/mean) was used to characterize the heterogeneity of perfusion during different postural and gravitational conditions.

(1)

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Physiological measures. All intravascular pressure, Paw, and Pes varied with the gravitational force. The time lag between changes in gravity and changes in the measured pressures (Fig. 2) was small relative to the time scale of interest. The physiological parameters reported in Table 2 were averaged over the final 7-9 s of each postural and gravitational condition across all parabolas. The esophageal balloon was inadvertently overinflated in all animals, causing artificially high Pes measurements. Despite this error, the difference in Pes between postures and gravitational conditions is accurate. In all animals and in all postures, central venous pressure, Paw, and Pes decreased significantly during the transition from 1.8 to 0 G and increased as the gravitational force increased (Table 2). Paw decreased by ~0.5 cmH₂O for each step drop in gravitational condition. Changes in systemic arterial pressure and pulmonary arterial pressure differed from this pattern. Systemic arterial pressure decreased by ~5 mmHg during weightlessness in the prone posture. Pulmonary arterial pressure decreased ~4 cmH₂O with each step decrease in the gravitational force when the animals were prone but did not change significantly when they were supine. Heart rate did not change significantly. Cardiac output varied significantly with both posture and gravitational condition. Compared with its position during weightlessness, the esophagus did not move significantly when the animals were prone. When supine, the esophagus moved an average of 0.2 and 0.3 cm toward the spine during 1- and 1.8-G conditions, respectively.

Gas exchange. Although alterations in gravitational forces were too short to allow gas exchange to reach a steady state, significant differences in gas exchange were observed between gravitational conditions. Table 3 presents the blood-gas data and the calculated alveolar-arterial O₂ difference (A-aDO₂) for each posture and gravitational condition. The A-aDO₂ was calculated for each postural and gravitational condition with the use of the alveolar gas equation (2), a respiratory quotient of 0.8, and the average cabin pressure during a given gravitational condition. A-aDO₂ increased significantly in both postures with increasing gravitational force. A significant difference was not observed between postures. The lowest arterial PO₂ across all animals in all postures and gravitational conditions was 83 Torr.

Regional blood flow. On average, the CV of perfusion was 0.67, 0.64, and 0.66 in the prone posture during 0-, 1-, and 1.8-G conditions, respectively. The CV of perfusion averaged 0.71, 0.71, and 0.77 in the supine posture during 0-, 1-, and 1.8-G conditions, respectively. There was a significant interaction between posture and gravitational force. Perfusion heterogeneity was significantly less in the prone vs. supine posture during 1- and 1.8-G conditions. Perfusion heterogeneity was not significantly different between postures during weightlessness.

View larger version (48K): [in this window] [in a new window]   Fig. 3.   Repeat measures of regional perfusion in 1 animal. Two different colors of microspheres were injected during similar postures and gravitational conditions, which were separated by 20 min and 10 parabolic loops. Deviation from a perfect correlation of 1.0 is due to temporal variability in perfusion and noise introduced by the microsphere method.

View larger version (56K): [in this window] [in a new window]   Fig. 4.   Regional perfusion in opposing postures during increased gravitational forces. Regional blood flow was determined in supine and prone postures during 1.8-G conditions in this animal. Despite changes in posture and increased gravitational forces, high-flow pieces remain high flow and low-flow pieces remain low flow. An effect of gravity can be demonstrated when spatial information is added to the plot. Dorsal lung pieces are colored blue, and ventral pieces are colored red. Note that the pieces that have the greatest change in hydrostatic gradient when rotated between postures have the greatest change in perfusion.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Vertical distribution of blood flow during 0-, 1-, and 1.8-G conditions in 1 animal in the supine posture. Slopes were determined from a least squares linear fit to blood flow as a function of height up the lung. We chose to exclude the traditional zone 4 region and fit only points with increasing blood flow down the lung. Values are means ± SE. Dependent and independent variable axes have been interchanged so that the formats are similar to previously published plots. Vertical gradient of perfusion was steeper at higher gravitational forces. Note that the general pattern of blood flow increasing and then decreasing down the lung (zones 3 and 4) persists during weightlessness.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This is the first study to measure regional pulmonary perfusion with microspheres during weightlessness and increased gravitational force. The microsphere method provides high spatial resolution of regional blood flow during different postures and gravitational forces within the same animal. The important findings of this study are that both gravity and the geometry of the pulmonary vascular tree determine regional blood flow. The vascular structure, however, is the primary determinant of local perfusion in these animals in both supine and prone postures.

The microsphere method for measuring regional organ perfusion is an established and well-accepted technique (23). It is based on the fundamental principle that labeled microspheres lodge within capillary beds in proportion to local blood flow. The microsphere method has been validated in the lung with the use of both radiolabeled red blood cells (3) and a "molecular microsphere" (15). A basic tenet is that injected microspheres are inert and do not alter local flow or resistance. It is argued that microspheres occlude only a fraction of capillaries and do not, therefore, significantly change local vascular resistance. However, there must be an upper limit to the number of microspheres that can be injected without causing hemodynamic changes. Pulmonary artery pressures and cardiac outputs were unchanged over the course of this study, suggesting that pulmonary vascular resistance was not altered by a cumulative dose of 10 million microspheres. Unfortunately, there is no literature in large animals to support the assumption that injecting microspheres does not alter vascular resistance. Using a pump-perfused rat lung with a fully dilated circulation, we have recently shown that pulmonary vascular resistance increases 0.8% for every 100,000 microspheres injected (9). Extrapolating this data from 0.25-kg rats to 25-kg pigs suggests that more than 10 million microspheres would need to be injected to raise vascular resistance 1%. In a lung in which the vasculature is not fully dilated, this increased resistance can be easily offset by vascular recruitment.

Our greatest concern in this study was that the pulmonary circulation would not have time to establish a steady state before microspheres were injected. The lack of significant lag between central vascular pressures and change in gravitational forces shows that the compartmental shifts of blood volume are remarkably rapid. Microspheres, however, lodge in the capillary bed, and we did not know how rapidly the microspheres would transit to their final wedged positions after abrupt changes in gravitational conditions. To determine the temporal response of the pulmonary capillary bed to rapid changes in flow, we used a pump-perfused lung preparation (14). We found that pulmonary capillary recruitment reached a new steady state within 4 s of a step change in flow. In the present study, microspheres were injected 5 s after the onset of a new gravitational condition. Given that steady state is reached in the peripheral pulmonary capillary bed in <4 s, we are confident that pulmonary blood flow to small lung pieces can reach a new steady state before microsphere injections.

A secondary concern was that the transit time from right atrium to the capillary bed would be too long to allow all of the microspheres to lodge before the onset of a new gravitational state. We estimated the mean transit time in our animals by dividing the pulmonary arterial blood volume by the cardiac output. Pulmonary blood volume is ~10% of total circulating blood volume, and about one-quarter of this resides in the pulmonary artery (6). Pigs have a circulating blood volume of 76 ml/kg (5), and we estimated the volume of blood in the pulmonary arteries of our pigs to be 50 ml. With a cardiac output of 3 liter/min, the mean transit time from the main pulmonary artery to the capillary bed will be 1 s. If four cardiac cycles are needed to wash microspheres out of the right ventricle, we estimate that 4-5 s are required for the majority of microspheres to reach the capillary bed following injection just proximal to the right atrium. Because of the heterogeneity of transit times in the lung (4), transit times to low-flow regions may be up to 4-5 s. Hence, 8-9 s may be required for the last microspheres to lodge in the lowest flow regions. All of our microsphere injections were completed more than 10 s before the next change in gravitational force. From this, we conclude that regional perfusion was determined during a constant gravitational force.

Our data on regional blood flow fit with concepts of pulmonary perfusion that have been evolving since Harvey demonstrated that the entire output of the right ventricle flowed through the lungs and Malpighi described the pulmonary microcirculation in 1661. Only 40 years ago, lung function was thought to be uniformly distributed. The first group to measure regional lung function in the early 1960s was understandably "surprised" to find a vertical distribution of ventilation and perfusion (26). Using methods with higher spatial resolution in the next decade, Greenleaf et al. (12) and Reed and Wood (21) uncovered more perfusion heterogeneity within isogravitational planes. These observations were later confirmed by Nicolaysen and colleagues (18). Even higher spatial resolution methods demonstrated that the spatial distribution of perfusion has an underlying pattern within isogravitational planes that was not random (7) and was constant over days (10). Most recently, studies demonstrated that perfusion remains heterogeneous during prolonged weightlessness (20).

One other study has measured perfusion heterogeneity during transient weightlessness (16). Amplitudes of cardiac oscillations were measured in exhaled respiratory gases from seated humans during parabolic flight. The amplitudes of the oscillations were nearly abolished during weightlessness compared with normal gravity. The authors concluded that virtually all the topographical inequality of blood flow seen under 1-G conditions was abolished during short periods of 0 G. These conclusions differ significantly from our current findings. Although we used supine and prone animals instead of upright humans, it is difficult to ascribe the large discrepancies to postural or species variation. Another possibility is that the exhaled gas method underestimated the true heterogeneity of pulmonary perfusion.

Recent studies performed on the NASA Spacelab Life Sciences-1 (SLS-1) (20) measured the amplitude of exhaled CO₂ from human subjects during prolonged weightlessness and normal gravity. The amplitudes of the cardiac oscillations decreased by 40% during weightlessness compared with at 1 G. The authors concluded that the decrease in amplitude of the cardiac oscillation may have been due to the absence of gravity and that the residual cardiac oscillations indicated persistent perfusion heterogeneity. These results agree more with our current findings. The gravitational influence may be larger in upright humans compared with supine and prone quadrupeds. In addition, measurement of exhaled CO₂ oscillations may underestimate perfusion heterogeneity. It is therefore difficult to compare the two studies quantitatively.

We think that our current data, data from studies performed in the 1960s, and data from SLS-1 can all be explained by one concept. In this model, regional pulmonary perfusion is determined primarily by the geometry of the vascular tree and the effects of gravity are superimposed on this underlying structure. When the spatial resolution of our current data is combined into isogravitational planes comparable to those of previous studies (27) (Fig. 5), they look very similar. Without the increased resolution of our measurements, it is not possible to observe the underlying pattern of perfusion heterogeneity. Data from humans in prolonged weightlessness demonstrated that, although perfusion heterogeneity decreased compared with normal gravitational conditions, the majority of perfusion heterogeneity persisted (20). The importance of vascular geometry does not preclude the waterfall effects first proposed by Riley et al. (22) and later conceptualized as zonal conditions by West and colleagues (28). We think that heterogeneities in vascular resistance can produce a range of zonal conditions within isogravitational planes (11) rather than being vertically stacked in the lung.

When averaged within isogravitational planes, blood flow increases and then decreases down the lung (Fig. 5). This pattern persists regardless of posture or gravitational force. Traditionally, the increase in perfusion is attributed to the hydrostatic gradient down the lung that distends dependent vessels, and increased interstitial pressures cause vessels to collapse in dependent regions (13). The observation that this pattern of increasing and then decreasing perfusion persists during weightlessness suggests that vertical pressure gradients may not explain this phenomenon in normal gravitational conditions. The relatively large extent of zone 4 conditions that we observed (Fig. 5) is probably related to the thin dorsal caudal lung in the pig, which is consistently compressed in the supine position under anesthesia. Although we hyperinflated the lungs before each microsphere injection, it may be that this lung region remained underinflated relative to its state in an awake, prone animal.

Our simple model of pulmonary blood flow uses only two determinants to explain perfusion heterogeneity: vascular structure and an imposed gravitational force. This model accounts for ~90% of the observed heterogeneity (Tables 4 and 5). The remaining 10% of perfusion heterogeneity not captured by the model may be caused by other factors, such as hypoxic pulmonary vasoconstriction and dependent lung compression. We attempted to minimize the effect of hypoxic pulmonary vasoconstriction by using a higher fraction of inspired O₂ and overventilating the animals to maintain a slight alkalosis. Transient lung compression due to pleural pressure gradients could not be avoided during microsphere injections. Atelectasis, which persisted across all gravitational states, would be erroneously considered vascular structure. We minimized this complication as much as possible by using recruitment maneuvers after each high-gravitational parabola.

Acknowledging the combined importance of vascular geometry and gravity on the regional distribution of pulmonary blood flow is an important advancement in our understanding of lung physiology. The shared effect of gravity can no longer be considered the mechanism matching regional ventilation and perfusion (19). Heterogeneous perfusion must somehow be matched by local ventilation within isogravitational planes. The mechanism by which postural changes improve gas exchange in patients with lung injury may be more complicated than we suspected (1). Studies exploring determinants of regional lung injury must consider the geometry of the pulmonary vascular tree in addition to hydrostatic gradients.