Fate of air emboli in the pulmonary circulation

The lung serves an important nonrespiratory function by trapping and excreting venous air emboli. The site of trapping and the mechanism of excretion, however, are uncertain. To observe the behavior of bubbles in the pulmonary circulation, we injected venous air emboli into anesthetized dogs and videotaped their elimination from the pulmonary microcirculation by using in vivo microscopy. Small intravenous bubbles lodged exclusively in pulmonary arterioles and were eliminated from that site. To determine whether the gas was dissolving into nearby blood and then was carried to the capillaries for excretion, the rate of bubble radius change was measured during nonperfused conditions produced by balloon occlusion of lobar blood flow and compared with perfused conditions. Bubble volume decreased at the same rate during perfused and nonperfused conditions and thus was independent of regional blood flow. Molecular diffusion of gas directly across the arteriolar wall into alveolar spaces was the most likely mechanism of elimination because calculations based on the Fick equation for molecular diffusion predict an elimination rate nearly identical with those observed experimentally.     

Ventilation-perfusion inhomogeneity increases gas uptake: theoretical modeling of gas exchange.

Ventilation-perfusion (VA/Q) inhomogeneity was modeled to measure its effect on gas exchange in the presence of inspired mixtures of two soluble gases using a two-compartment computer model. Theoretical studies involving a mixture of hypothetical gases with equal solubility in blood showed that the effect of increasing inhomogeneity of distributions of either ventilation or blood flow is to paradoxically increase uptake of the gas with the lowest overall uptake in relation to its inspired concentration. This phenomenon is explained by the concentrating effects that uptake of soluble gases exert on each other in low VA/Q compartments. Repeating this analysis for inspired mixtures of 30% O(2) and 70% nitrous oxide (N(2)O) confirmed that, during "steady-state" N(2)O anesthesia, uptake of N(2)O is predicted to paradoxically increase in the presence of worsening VA/Q inhomogeneity.     

The secretion of inert gas into the swim-bladder of fish

The composition of the gas mixture secreted into the swim-bladders of several species of fish has been determined in the mass spectrometer. The secreted gas differed greatly from the gas mixture breathed by the fish in the relative proportions of the chemically inert gases, argon, neon, helium, and nitrogen. Relative to nitrogen the proportion of the very soluble argon was increased and the proportions of the much less soluble neon and helium decreased. The composition of the secreted gas approaches the composition of the gas mixture dissolved in the tissue fluid. A theory of inert gas secretion is proposed. It is suggested that oxygen gas is actively secreted and evolved in the form of minute bubbles, that inert gases diffuse into these bubbles, and that the bubbles are passed into the swim-bladder carrying with them inert gases. Coupled to a preferential reabsorption of oxygen from the swim-bladder this mechanism can achieve high tensions of inert gas in the swim-bladder. The accumulation of nearly pure nitrogen in the swim-bladder of goldfish (Carassius auratus) is accomplished by the secretion of an oxygen-rich gas mixture followed by the reabsorption of oxygen.     

Continuous measurement of gas uptake and elimination in anesthetized patients using an extractable marker gas.

Measurement of pulmonary gas uptake and elimination is often performed, using nitrogen as marker gas to measure gas flow, by applying the Haldane transformation. Because of the inability to measure nitrogen with conventional equipment, measurement is difficult during inhalational anesthesia. A new method is described, which is compatible with any inspired gas mixture, in which fresh gas and exhaust gas flows are measured using carbon dioxide as an extractable marker gas. A system was tested in eight patients undergoing colonic surgery for automated measurement of uptake of oxygen, nitrous oxide, isoflurane, and elimination of carbon dioxide with this method. Its accuracy and precision were compared with simultaneous measurements made with the Haldane transformation and corrected for predicted nitrogen excretion by the lungs. Good agreement was obtained for measurement of uptake or elimination of all gases studied. Mean bias was -0.003 l/min for both oxygen and nitrous oxide uptake, -0.0002 l/min for isoflurane uptake, and 0.003 l/min for carbon dioxide elimination. Limits of agreement lay within 30% of the mean uptake rate for nitrous oxide, within 15% for oxygen, within 10% for isoflurane, and within 5% for carbon dioxide. The extractable marker gas method allows accurate and continuous measurement of gas uptake and elimination in an anesthetic breathing system with any inspired gas mixture.     

Response of pulmonary veins to increased intracranial pressure and pulmonary air embolization

Brain compression with subdural air causes pulmonary hypertension and noncardiogenic pulmonary edema (A. B. Malik, J. Appl. Physiol.: Respirat. Environ. Exercise Physiol. 42: 335-343, 1977). To see whether air emboli to the lungs rather than brain compression caused these findings, anesthetized dogs received intravenous air infusions, subdural air infusions, or brain compression from balloons inflated in the subdural space. Subdural air and intravenous air resulted in similar vascular responses. Pulmonary artery pressure (Ppa) increased 160% (P less than 0.01) and pulmonary venous pressure transiently rose 13 +/- 5 Torr (P less than 0.05) without an increase in left atrial pressure or cardiac output (Q). The end-tidal PCO2 fell 55% (P less than 0.01) and the postmortem weight of the lungs increased 55% (P less than 0.05). Brain compression with a subdural balloon instead of air only caused a 20% rise in Ppa and Q without pulmonary edema. Thus, pulmonary air emboli rather than brain compression accounts for the edema and pulmonary hypertension caused by subdural air. Catheters in pulmonary veins and the left atrium showed that air emboli cause transient pulmonary venous hypertension as well as a reproducible form of noncardiogenic pulmonary endema.     

Antagonism of membrane compression effects by high pressure gas mixtures in a phospholipid bilayer system

Binary mixtures of helium with nitrogen, xenon or nitrous oxide were applied to suspensions of phosphatidylcholine-cholesterol vesicles to determine those mixtures of lipid soluble gases which would exactly antagonize the membrane rigidifying effect of 100 ATA compression. A previous study has shown that the initial application of 100 ATA compression by gas produces a significant reduction in the fluidity of the phospholipid bilayer. However, as the high pressure gas dissolves into the lipid region it creates disorder and increases fluidity. Fluidity of the bilayer at equilibrium represents the sum of the compression-ordering and dissolved-gas disordering effects and is dependent on the gas/lipid partition coefficient of the particular gas. The beneficial effect of a narcotic gas added to Trimix mixtures to ameliorate HPNS in deep divers may be due to a balance of compression-ordering and solubility-disordering effects achieved within the nerve membrane. It is therefore valuable to determine those gas mixtures which achieve balance of these two effects and result in zero net change in phospholipid bilayer fluidity at an established pressure of 100 ATA. Binary mixtures of helium with 88% nitrogen, 3.8% xenon or 2.8% nitrous oxide resulted in zero net change in bilayer fluidity with our model system at 100 ATA. A graph of the percent of narcotic gas needed to produce zero net effect as a function of pressure, however, was nonlinear. This would suggest the ratio of gases in Trimix must be varied as a function of pressure. While the phosphatidylcholine-cholesterol bilayer is a good model for certain components of the nerve membrane, it does not allow for study of protein-lipid or gas-protein interactions. The data presented thus aid in our understanding of HPNS but are yet incomplete for precise use in predicting diving mixtures.     

Cardiac output by rebreathing in patients with cardiopulmonary diseases

Noninvasive estimates of cardiac output by rebreathing soluble gases (Qc) can be unreliable in patients with cardiopulmonary diseases because of uneven distribution of ventilation to lung gas volume and pulmonary blood flow. To evaluate this source of error, we compared rebreathing Qc with invasive measurements of cardiac output performed by indicator-dilution methods (COID) in 39 patients with cardiac or pulmonary diseases. In 16 patients with normal lung volumes and 1-s forced expiratory volumes (FEV1), Qc measured with acetylene [Qc(C2H2)] overestimated COID insignificantly by 2 +/- 9% (SD). In subjects with mild to moderate obstructive lung disease, Qc(C2H2) slightly overestimated COID by 6 +/- 15% (P = 0.11). In patients with restrictive disease or combined obstructive and restrictive disease, Qc(C2H2) underestimated COID significantly by 9 +/- 14% (P less than 0.04). The magnitude of the discrepancy between Qc and COID correlated with size of the volume rebreathed and an index of uneven ventilation calculated from helium mixing during rebreathing that determined a dead space to inspired volume ratio (VRD/VI). Rebreathing volumes less than 40% of the predicted FEV or VRD/VI of 0.4 or greater identified all subjects with a discrepancy between Qc(C2H2) and COID of 20% or greater.     

Factors determining temporal pattern of isobaric supersaturation

It is possible to produce a transient supersaturation or undersaturation in tissues and blood by sequentially breathing gases with different equilibration rates. If the ambient gas pressure is sufficiently high, the induced supersaturation can produce vascular bubbles. By means of the classical perfusion-dependent model of inert gas elimination, which assumes that the effects of diffusion are minimal, the magnitude of the total inert gas pressure can be predicted. If, however, the effects of diffusion cannot be ignored, the supersaturation could be substantially larger. This paper estimates the effects of diffusion in a Krogh cylinder on the supersaturation produced by suddenly changing the inert gas partial pressure in the blood. The results of these estimates indicate that diffusion plays a role in this transient supersaturation only in long Krogh cylinders with high blood flows. The effects of diffusion are further reduced by the finite time necessary to switch the inert gases in arterial blood. The conclusions are supported by experiments that measure vascular bubble production after a switch of the inert portion of the inspired gas. These experiments further show that the formation of vascular bubbles after such a switch cannot be entirely explained by the different diffusion constants of the gases used.     

Convective and diffusive gas transport in canine intrapulmonary airways.

The significance of convective and diffusive gas transport in the respiratory system was assessed from the response of combined inert gas and particle boluses inhaled into the conducting airways. Particles, considered as "nondiffusing gas," served as tracers for convection and two inert gases with widely different diffusive characteristics (He and SF6) as tracers for convection and diffusion. Six-milliliter boluses labeled with monodisperse di-2-ethylhexyl sebacate droplets of 0.86-microns aerodynamic diameter, 2% He, and 2% SF6 were inspired by three anesthetized mechanically ventilated beagle dogs to volumetric lung depths up to 170 ml. Mixing between inspired and residual air caused dispersion of the inspired bolus, which was quantified in terms of the bolus half-width. Dispersion of particles increased with increasing lung depth to which the boluses were inhaled. The increase followed a power law with exponents less than 0.5 (mean 0.39), indicating that the effect of convective mixing per unit volume was reduced with depth. Within the pulmonary dead space, the behavior of the inert gases He and SF6 was similar to that of the particles, suggesting that gas transport was almost solely due to convection. Beyond the dead space, dispersion of He and SF6 increased more rapidly than dispersion of particles, indicating that diffusion became significant. The gas and particle bolus technique offers a suitable approach to differential analysis of gas transport in intrapulmonary airways of lungs.     

Ventilation-perfusion distribution in normal subjects

Functional values of LogSD of the ventilation distribution (σ(V)) have been reported previously, but functional values of LogSD of the perfusion distribution (σ(q)) and the coefficient of correlation between ventilation and perfusion (ρ) have not been measured in humans. Here, we report values for σ(V), σ(q), and ρ obtained from wash-in data for three gases, helium and two soluble gases, acetylene and dimethyl ether. Normal subjects inspired gas containing the test gases, and the concentrations of the gases at end-expiration during the first 10 breaths were measured with the subjects at rest and at increasing levels of exercise. The regional distribution of ventilation and perfusion was described by a bivariate log-normal distribution with parameters σ(V), σ(q), and ρ, and these parameters were evaluated by matching the values of expired gas concentrations calculated for this distribution to the measured values. Values of cardiac output and LogSD ventilation/perfusion (Va/Q) were obtained. At rest, σ(q) is high (1.08 ± 0.12). With the onset of ventilation, σ(q) decreases to 0.85 ± 0.09 but remains higher than σ(V) (0.43 ± 0.09) at all exercise levels. Rho increases to 0.87 ± 0.07, and the value of LogSD Va/Q for light and moderate exercise is primarily the result of the difference between the magnitudes of σ(q) and σ(V). With known values for the parameters, the bivariate distribution describes the comprehensive distribution of ventilation and perfusion that underlies the distribution of the Va/Q ratio.     

Ventilation distribution in rats: Part I - The effect of gas composition as measured with electrical impedance tomography

ABSTRACT: The measurement of ventilation distribution is currently performed using inhaled tracer gases for multiple breath inhalation studies or imaging techniques to quantify spatial gas distribution. Most tracer gases used for these studies have properties different from that of air. The effect of gas density on regional ventilation distribution has not been studied. This study aimed to measure the effect of gas density on regional ventilation distribution. METHODS: Ventilation distribution was measured in seven rats using electrical impedance tomography (EIT) in supine, prone, left and right lateral positions while being mechanically ventilated with either air, heliox (30% oxygen, 70% helium) or sulfur hexafluoride (20% SF6, 20% oxygen, 60% air). The effect of gas density on regional ventilation distribution was assessed. RESULTS: Gas density did not impact on regional ventilation distribution. The non-dependent lung was better ventilated in all four body positions. Gas density had no further impact on regional filling characteristics. The filling characteristics followed an anatomical pattern with the anterior and left lung showing a greater impedance change during the initial phase of the inspiration. CONCLUSION: It was shown that gas density did not impact on convection dependent ventilation distribution in rats measured with EIT.     

The Role of Cardiac Shunts in the Regulation of Arterial Blood Gases

SYNOPSIS. The pulmonary and systemic circulations are not completelyseparated in reptiles and amphibians, so oxygen-rich blood returningfrom the lungs can mix with oxygen-poor blood returning fromthe systemic circuit (cardiac shunts). In these animals, thearterial blood gas composition is determined by both lung ventilationand the cardiac shunt. Therefore, changes in cardiac shuntingpatterns may participate actively in the regulation of arterialblood gases. In turtles the cardiac shunt pattern changes independentlyof ventilation and the cardiac R-L shunt (pulmonary bypass ofsystemic venous blood) is reduced under circumstances wherethe demands on efficient gas exchange are high (hypoxia, hypoxemiaor exercise). We propose, therefore, that the size of cardiacshunts is regulated independently of ventilation and hypothesizethat there exist at least two groups of peripheral chemoreceptorswith different reflex roles.     

Determinants of arterial gas embolism after scuba diving

Scuba diving is associated with breathing gas at increased pressure, which often leads to tissue gas supersaturation during ascent and the formation of venous gas emboli (VGE). VGE crossover to systemic arteries (arterialization), mostly through the patent foramen ovale, has been implicated in various diving-related pathologies. Since recent research has shown that arterializations frequently occur in the absence of cardiac septal defects, our aim was to investigate the mechanisms responsible for these events. Divers who tested negative for patent foramen ovale were subjected to laboratory testing where agitated saline contrast bubbles were injected in the cubital vein at rest and exercise. The individual propensity for transpulmonary bubble passage was evaluated echocardiographically. The same subjects performed a standard air dive followed by an echosonographic assessment of VGE generation (graded on a scale of 0-5) and distribution. Twenty-three of thirty-four subjects allowed the transpulmonary passage of saline contrast bubbles in the laboratory at rest or after a mild/moderate exercise, and nine of them arterialized after a field dive. All subjects with postdive arterialization had bubble loads reaching or exceeding grade 4B in the right heart. In individuals without transpulmonary passage of saline contrast bubbles, injected either at rest or after an exercise bout, no postdive arterialization was detected. Therefore, postdive VGE arterialization occurs in subjects that meet two criteria: 1) transpulmonary shunting of contrast bubbles at rest or at mild/moderate exercise and 2) VGE generation after a dive reaches the threshold grade. These findings may represent a novel concept in approach to diving, where diving routines will be tailored individually.     

Effect of high-frequency ventilation on gas exchange and pulmonary vascular resistance in lambs

We studied the effects of conventional mechanical ventilation (CMV) (15 ml/kg tidal volume delivered at 18-25 breaths/min) and high-frequency oscillatory ventilation (HFOV) (less than or equal to 2 ml/kg delivered at 10 Hz) on pulmonary hemodynamics and gas exchange during ambient air breathing and hypoxic gas breathing in 10 4-day-old lambs. After instrumentation and randomization to either HFOV or CMV the animals breathed first ambient air and then hypoxic gas (inspired O2 fraction = 0.13) for 20 min. The mode of ventilation was then changed, and the normoxic and hypoxic gas challenges were repeated. The multiple inert gas elimination technique was utilized to assess gas exchange. There was a significant increase with HFOV in mean pulmonary arterial pressure (Ppa) (20.1 +/- 4.2 vs. 22 +/- 3.8 Torr, CMV vs. HFOV, P less than 0.05) during ambient air breathing. During hypoxic gas breathing Ppa was also greater with HFOV than with CMV (29.5 +/- 5.7 vs. 34 +/- 3.1 Torr, CMV vs. HFOV, P less than 0.05). HFOV reduced pulmonary blood flow (Qp) during ambient air breathing (0.33 +/- 0.11 vs. 0.28 +/- 0.09 l . kg-1 . min-1, CMV vs. HFOV, P less than 0.05) and during hypoxic gas breathing (0.38 +/- 0.11 vs. 0.29 +/- 0.09 l . kg-1 . min-1, P less than 0.05). There was no significant difference in calculated venous admixture for sulfur hexafluoride or in the index of low ventilation-perfusion lung regions with HFOV compared with CMV.(ABSTRACT TRUNCATED AT 250 WORDS)     

Intrapulmonary Co2 receptors and ventilatory response to lung Co2 loading

The possible role of intrapulmonary CO2 receptors (IPC) in arterial CO2 partial pressure (PaCO2) homeostasis was investigated by comparing the arterial blood gas and ventilatory responses to CO2 loading via the inspired gas and via the venous blood. Adult male Pekin ducks were decerebrated 1 wk prior to an experiment. Venous CO2 loading was accomplished with a venovenous extracorporeal blood circuit that included a silicone-membrane blood oxygenator. The protocol randomized four states: control (no loading), venous CO2 loading, inspired CO2 loading, and venous CO2 unloading. Intravenous and inspired loading both resulted in hypercapnic hyperpnea. Comparison of the ventilatory sensitivity (delta VE/delta PaCO2) showed no significant difference between the two loading regimes. Likewise, venous CO2 unloading led to a significant hypocapnic hypopnea. Sensitivity to changes in PaCO2 could explain the response of ventilation under these conditions. The ventilatory pattern, however, was differentially sensitive to the route of CO2 loading; inspired CO2 resulted in slower deeper breathing than venous loading. It is concluded that IPC play a minor role in adjusting ventilation to match changes in pulmonary CO2 flux but rather are involved in pattern determination.     

The compression-ordering and solubility-disordering effects of high pressure gases on phospholipid bilayers.

Inert gases at high pressure may compress and dissolve in tissue of intact organism to result in narcosis, reversal of the effects of anesthetic agents or hyperexcitability. The effects of 51 and 102 atm of helium, hydrogen, nitrogen, argon, xenon and nitrous oxide on the molecular motion of nitroxide spin-labeled phospholipid-cholesterol bilayers were measured by electron paramagnetic resonance (EPR) techniques. Immediately, application of high pressures of all gases decreased the molecular motion of the fatty acid chains of the membrane phospholipids; the magnitude of ordering was linearly related to the amount of pressure applied. The second effect was an increase in molecular motion of the fatty acid chains which appeared more slowly due to the slow gas diffusion through the column of lipid dispersion. The magnitude of disorder of the phospholipid membrane at equilibrium correlated with the known lipid solubilities of the gases in olive oil as well as with the anesthetic potency of all the gases except xenon. The environment of the spin label became less polar as the gases diffused into the bilayer. The present studies in the phospholipid model membrane show that the net effects of high pressure gases in the lipid phase consist of an initial ordering of the membrane by compression opposed by the ability of the gas molecules to diffuse and dissolve in the lipid bilayers and disorder them. It is thus suggested that the resultant perturbations of the membrane lipid fluidity by high pressure gases may subsequently be transmitted to membrane-bound protein to result in changes that may be associated, in part, with the diverse effects of anesthesia and of the high pressure nervous syndrome (HPNS) observed in deep-sea divers. The model system may be useful in developing gas mixtures which minimize HPNS.     

The extracranial venous system in the heads of beaked whales,with implications on diving physiology and pathogenesis

Beaked whales are a poorly known but diverse group of whales that have received considerable attention due to strandings that have been temporally and spatially associated with naval sonar deployment. Postmortem studies on stranded carcasses have revealed lesions consistent with decompression sickness, including intravascular gas and fat emboli. These findings have been supported by analyses of intravascular gas emboli showing composition dominated by nitrogen gas. To increase our understanding of the pathophysiology of nitrogen bubble formation and intravascular embolization, we examined the gross and microscopic anatomy of the venous system in the head of beaked whales. Since the potential sources of intravascular fat and gas emboli were of greatest interest, focus was placed on the acoustic fat bodies and pneumatic accessory sinus system. Herein, we describe intimate arteriovenous associations with specialized adipose depots and air sinuses in beaked whales. These vascular structures comprise an extensive network of thin‐walled vessels with a large surface area, which is likely to facilitate exchange of nitrogen gas and may, therefore, form anatomic regions that may be important in physiological management of diving gases. These structures may also be vulnerable to pathologic introduction of emboli into the vascular system. Expansive, thin‐walled venous lakes are found within the pterygoid region, which suggest the potential for nitrogen exchange as well as for compensation of middle‐ear pressures during descent on a dive. These findings warrant further research into the structure and function of this morphology as it relates to normal and pathologic physiology. J. Morphol. 277:34–64, 2016. © 2015 Wiley Periodicals, Inc.     

Effect of aminophylline on transpulmonary passage of venous air emboli in pigs.

Aminophylline has been shown to dramatically reduce the filtering capacity of the lung in dogs during venous air embolism. Similarities have been pointed out between the cardiovascular and respiratory systems of the pig and of humans. We therefore wanted to find out whether aminophylline also modifies the transpulmonary spillover of microbubbles to the arterial circulation of the pig. Twenty-eight pigs were anesthetized with pentobarbital sodium and mechanically ventilated. Aminophylline was injected intravenously into 10 of the pigs before the introduction of air bubbles into the right ventricle, while the other 18 pigs served as controls. A transesophageal echocardiographic probe was used to detect eventual air bubbles in the left atrium or in the aorta. Pigs received either air infusion, at rates varying from 0.05 to 0.20 ml.kg-1.min-1, or calibrated microbubbles, 5-300 microns diam. We found that aminophylline-treated pigs did not show any change in spillover incidence compared with controls. Furthermore, in both groups the spillover during continuous air infusion seemed to be a preterminal event, because the pigs had very low arterial pressure when arterial bubbles were observed. Finally, there was an increase in mean pulmonary arterial pressure from 18 +/- 3.4 to 26 +/- 2.2 (SD) mmHg (n = 4, P less than 0.01) in aminophylline-treated pigs after a bolus injection of microbubbles (less than or equal to 50 microns, total volume less than 0.5 ml). Our results suggest that aminophylline does not modify the transpulmonary passage of microbubbles in this porcine model. In addition, it would seem that the pulmonary circulation of the pig is sensitive to very small volumes of air, when injected as microbubbles.     

Carbon dioxide-oxygen relationships in gas exchange of animals. In memory of Hermann Rahn.

In external gas exchange of vertebrates, behavior of the respiratory gases CO2 and O2 can in many cases adequately be explained by the different physico-chemical properties of the gases, including solubility, chemical combination in blood and tissue, and diffusivity. In particular, the differences in behavior between CO2 and O2 are often of particular relevance. This is demonstrated on a number of examples of gas exchange mechanisms in vertebrates, including (1) exchange ratio after changes in ventilation, (2) local variations of pulmonary ventilation/perfusion ratio, (3) absorption of gas from gas pockets, (4) water vs. air breathing, (5) multimodal breathing, (6) skin breathing, (7) gas exchange of avian eggs, (8) anomalous gas/blood CO2 equilibration, (9) blood/gas CO2 equilibration in avian lungs, (10) pulmonary diffusing capacity, (11) blood/water CO2 equilibration in fish gills, (12) deposition of gas into fish swim bladder.     

Mechanism of stimulation of pulmonary prostacyclin synthesis at birth

In order to investigate the mechanism behind ventilation-induced pulmonary prostacyclin production at birth, chloralose anesthetized, exteriorized, fetal lambs were ventilated with a gas mixture that did not change blood gases (fetal gas) and unventilated fetal lungs were perfused with blood containing increased O2 and decreased CO2. Ventilation with fetal gas (3%O2, 5%CO2) increased net pulmonary prostacyclin (as 6-keto-PGF1 alpha) production from -5.1 +/- 4.4 to +12.6 +/- 7.6 ng/kg X min. When ventilation was stopped, net pulmonary prostacyclin production returned to nondetectable levels. Ventilation with gas mixtures which increased pulmonary venous PO2 and decreased PCO2 also stimulated pulmonary prostacyclin production, but did not have greater effects than did ventilation with fetal gas. In order to determine if increasing PO2 or decreasing PCO2 could stimulate pulmonary prostacyclin production independently from ventilation, unventilated fetal lamb lungs were perfused with blood that had PO2 and PCO2 similar to fetal blood, blood with elevated O2, and blood that had PO2 and PCO2 values similar to arterial blood of newborn animals. Neither increased O2 nor decreased CO2 in the blood perfusing the lungs stimulated pulmonary prostacyclin synthesis. We conclude that the mechanism responsible for the stimulation of pulmonary prostacyclin production with the onset of ventilation at birth is tissue stress during establishment of gaseous ventilation and rhythmic ventilation.