INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Neurogenic pulmonary edema (NPE) is a form of edema that may develop rapidly after such central nervous system (CNS) insults as head injury, subarachnoid hemorrhage, or epileptic seizures (1). These stimuli may produce massive sympathetic nervous system (SNS) activation, leading to extreme, but transient, episodes of systemic and pulmonary hypertension, the latter resulting in edema. In a recent study, we followed the time course of the changes in extravascular lung water that occurred in a canine model of NPE (4) in which the SNS was centrally activated by the intracisternal injection of veratrine and observed that the edema appeared to recover fairly rapidly (5). This observation is consistent with case reports that indicate that the edema resolves relatively quickly in many patients who survive the initial CNS insult (1).

The results of studies conducted during the last decade and a half examining the mechanisms of alveolar epithelial vectorial sodium and water transport have provided important insight into how excess fluid is cleared from the alveolar spaces and provide a basis for developing an understanding of the rapid resolution of NPE observed in both the human and dog. In this regard, excess water is believed to be osmotically cleared from the alveolar spaces by a mechanism involving sodium transport across the alveolar epithelium (6). Specifically, sodium is thought to enter the alveolar epithelial type II cell through multiple specialized pathways located in the apical membrane and then pumped out of the basolateral side by the enzyme Na⁺-K⁺-adenosinetriphosphatase (Na⁺-K⁺-ATPase). Of particular interest are observations indicating that ₂-adrenergic agonists (e.g., terbutaline, epinephrine) can increase the rate of sodium and water reabsorption from fluid-filled lungs of most species (including the human) and the rate of sodium and water movement across alveolar epithelial cell monolayers (6, 7). These observations suggest that the lung possesses intrinsic regulatory mechanisms that allow it to maintain its level of hydration under normal conditions and to provide for the accelerated removal of fluid when required (8).

The ability of ₂-adrenergic agonists to increase alveolar liquid clearance (ALC), coupled with our previous observations that sustained, large increases in plasma epinephrine concentration develop in the canine veratrine NPE model (9), suggested to us that ALC and, consequently, the rate of recovery from NPE might be accelerated by endogenous epinephrine. To evaluate this possibility, in this study, we initially determined if ALC was increased in anesthetized dogs in which the SNS had been massively activated by the intracisternal injection of veratrine and found that ALC was elevated 116% over baseline values. This observation indicated that the rate of fluid removal from the airspaces after SNS activation was accelerated by an endogenous mechanism. Accordingly, we designed a series of further experiments to determine if the increased ALC was mediated by ₂-adrenoceptors, epinephrine released from the adrenal glands, and involved an increase in alveolar epithelial sodium transport.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experiments were performed on 68 dogs of mixed breed and sex (21.1 ± 3.8 kg [SD]). In early experiments, which were designed to validate our technique (an initial control group and animals administered either terbutaline or terbutaline plus the ₂-adrenoceptor antagonist ICI 118,551), the dogs were anesthetized with thiamylal sodium (18 mg/kg, intravenously) followed by -chloralose (50 mg/kg, intravenously). Although terbutaline and ICI 118,551 changed ALC in a predictable manner (see RESULTS), we noted that ALC in the control group was significantly lower than that previously observed in both awake and pentobarbital anesthetized dogs (10) and in isolated nonperfused dog lung lobes (11). This comparison suggested that the thiamylal sodium/-chloralose anesthetic combination may have depressed the absolute level of ALC in all three groups. Accordingly, all subsequent experiments (including a new control group) were conducted under pentobarbital sodium anesthesia (30 mg/kg, intravenously).

After the dogs were anesthetized, they were intubated and ventilated with a piston respirator. Catheters were placed in the right femoral and pulmonary (flow-directed thermodilution catheter) artery to monitor arterial (Pa), pulmonary arterial (Ppa), and wedge (Ppaw) pressures, and in the right femoral vein for intravenous infusions. Additional large bore catheters were placed in the right carotid artery and jugular vein to be used as described subsequently. The animals were ventilated with 30% O₂ at an average frequency of 8.5 ± 1.2 breaths/min and tidal volume of 449 ± 67 ml. The average peak inspiratory pressure was 8.9 ± 1.1 mm Hg under these conditions. Arterial blood gases were analyzed using a Radiometer system and were as follows under baseline conditions: PO₂, 130.0 ± 17.3 mm Hg; PCO₂, 35.9 ± 3.4 mm Hg; and pH, 7.40 (range 7.34 to 7.51). A polyethylene catheter (3 mm, inner diameter) was placed in a lower lung lobe airway through a port in the endotracheal tube to allow us to instill plasma into the airspaces. Body temperature was maintained by a water-perfused heating pad placed beneath the animal. The dog's blood was heparinized (1,000 U/kg).

Determination of Alveolar Liquid Clearance

ALC was determined using the method of Berthiaume and coworkers (10, 12). With this approach, increases occurring over time in the protein concentration of the plasma instilled in the alveoli represent the absorption of fluid from the airspaces. Because the rate of alveolar protein clearance is comparatively very slow (~ 1%/h) (10, 12), the rate of fluid absorption can be calculated by mass balance. For each experiment, ~ 7 ml/kg arterial blood was drawn from the animal, centrifuged, and replaced with an equal volume of 6% dextran. For each experiment, 3 ml/kg plasma was instilled in a lower lung lobe through a port in the endotracheal tube.

The volume of instilled plasma remaining in the lung at the end of the experiment (V_f) was calculated by mass balance (10, 12):

(1)

where V_i is the initial instilled volume, Pr_i is the initial protein concentration, and Pr_f is the final protein concentration. ALC, expressed as a percentage of the volume instilled, was calculated as (13):

(2)

where Fw_i and Fw_f are, respectively, the water fraction of the initial plasma instillate and that of the instillate at the end of the experiment. These were determined gravimetrically.

Protein concentrations were determined by refractometry (American Optical, Buffalo, NY). We have previously evaluated the ability of the American Optical instrument to accurately measure the increases in alveolar instillate protein concentration that occur as fluid is absorbed from the airspaces under both baseline conditions and when ALC was accelerated by either terbutaline administration or massive SNS activation (11). Briefly, the refractometer protein determination was found to correlate strongly (r = 0.99, r² = 0.98, p < 0.001) with the protein concentration measured using the Waddell method (14).

Comparison of Residual Alveolar Water and Excess Lung Water

In each experiment, the residual alveolar water content of the lungs was compared with the total excess lung water content. Residual alveolar water (the water volume of the instilled plasma remaining in the airspaces) was calculated as the product of V_f and Fw_f. Excess lung water (the water volume of the instilled plasma remaining in the lung irrespective of the anatomic site) was determined in the following manner. At the end of the experiment, the lobe in which plasma had been instilled (wet lobe) was removed for the gravimetric determination of extravascular lung water content (EVLW) using the method of Pearce and coworkers (15). EVLW of the normal (dry) right upper lobe (RUL) was also determined. Excess lung water was determined by subtracting the calculated EVLW of the wet lobe prior to plasma instillation (i.e., under baseline conditions) from the measured EVLW at the end of the experiment (10, 12). The baseline EVLW of the wet lobe was calculated by multiplying the EVLW/blood-free dry weight ratio of the RUL times the dry weight of the wet lobe. The dry weight of the wet lobe was corrected for the dry weight of the plasma instilled in the wet lobe. This method is based on the assumption that under baseline conditions, the EVLW/dry weight ratio is identical in the two lobes. We have previously measured this ratio in separated lung lobes from the same animal and found this assumption to be valid.

Experimental Groups and Specific Protocols

The following groups of animals were studied:

Veratrine group. After the preparation had stabilized, baseline vascular pressure measurements and thermodilution cardiac outputs (American Edwards Laboratories, Santa Ana, CA) were made, and a 1-ml arterial blood sample was drawn for the determination of blood gases and plasma protein concentration. An additional 5-ml arterial sample was drawn for the analysis of epinephrine and norepinephrine concentrations by high-performance liquid chromatography (HPLC) as previously described (9). The plasma was then instilled in the lung, and additional hemodynamic measurements were made after 5 min. In six dogs, veratrine (100 µg/kg; Sigma Chemical, St. Louis, MO) was injected intracisternally as previously described (4, 5, 9). Subsequent hemodynamic measurements were made 10 min after veratrine administration (the time at which we have previously observed plasma catecholamine concentrations to be at their highest levels [9]), at 30 min, and then again at 30-min intervals for a period of 4 h. Blood gases were determined 10 min after veratrine injection and then at hourly intervals. Sodium bicarbonate was administered as necessary to maintain arterial pH within a normal range. Plasma protein concentration was determined at hourly intervals. Plasma catecholamine determinations were made at 10 min, 2 h, and 4 h after veratrine administration. At the end of 4 h, the animal was killed with an overdose of pentobarbital sodium, and the lungs were removed. At this time, a well-mixed alveolar fluid sample was collected for both protein and catecholamine analyses to be compared with similar measurements made on the initial instillate.

To prevent the development of pulmonary hypertension and resulting edema (4, 5) that typically occurs in animals receiving veratrine (which would confound the mass balance ALC determinations), the carotid arterial catheter was opened in these studies as soon as Ppa started to rise after veratrine administration, thus allowing blood to empty into a stirred, heated reservoir. After Ppa started to return to normal (30 to 60 min after veratrine injection), the blood was returned to the jugular vein. This procedure was also required for the following groups of animals (see below): veratrine + ICI 118,551; veratrine + amiloride, epinephrine, and epinephrine + ICI 118,551. The volume of blood removed in this manner averaged 282 ± 157 (SD) ml. The effectiveness of the controlled hemorrhage procedure in preventing the development of edema was verified in each experiment in which the procedure was used by determining EVLW of the RUL. In these experiments (n = 27), RUL EVLW averaged 3.53 ± 0.26 (SD) g H₂O/g blood-free dry weight. This value was within the range that we previously observed in normal dog lungs (4). For all of the following groups, blood sampling, biochemical analyses, and hemodynamic measurements were made at the same times described above for the veratrine group.

Control group. In six animals, plasma was instilled into a lower lung lobe as described previously, and baseline rates of ALC were determined after 4 h.

Veratrine + adrenalectomy group. Because we have previously shown that adrenalectomy eliminates the extreme increases in plasma epinephrine concentration that occur after veratrine administration (9), we evaluated the possibility that a circulating agent released from the adrenal glands (most probably epinephrine) increases ALC after SNS activation by determining if veratrine increased ALC in six acutely adrenalectomized dogs. The animals were adrenalectomized through incisions in the right and left flanks. To minimize the potential for abdominal bleeding, the animals' blood was not heparinized in the adrenalectomized group. Two additional adrenalectomized dogs were studied under baseline conditions to determine if adrenalectomy affected baseline clearance rates.

Veratrine + ICI 118,551 group. These experiments (n = 6) were designed to determine the ability of the highly specific ₂-adrenoceptor antagonist ICI 118,551 (Tocris Cookson, St. Louis, MO) to inhibit the increased ALC produced by veratrine. ICI 118,551 was initially administered as a bolus (100 µg/kg, intravenously) 10 min before instilling the plasma and followed by an intravenous infusion (1.0 µg/kg/ min) for the duration of the experiment (16). In addition to intravenous administration, the plasma instilled in the lungs also contained 20 µg/ml ICI 118,551. Three additional animals were studied to determine if ICI 118,551 altered baseline ALC.

Veratrine + amiloride group. In five dogs, the sodium channel blocker amiloride (10⁴ M; Sigma) was added to the instilled plasma to determine if the increased ALC was mediated by an increase in alveolar epithelial sodium transport. Three additional dogs were studied to determine the effect of amiloride on baseline ALC.

Terbutaline group. In five dogs, terbutaline, a ₂-adrenergic agonist (Sigma), was dissolved (10⁵ M) in the plasma that was instilled in the lungs to confirm the ability of a ₂-adrenergic agonist to increase ALC in our preparation and to provide a basis of comparison to allow us to evaluate the ability of ICI 118,551 to block ₂-adrenoceptors. Determinations of ALC made in animals administered terbutaline or terbutaline + ICI 118,551 (see below) were compared with those made in a separate group of control animals (n = 8) anesthetized with the same anesthetic (thiamylal sodium and -chloralose).

Terbutaline + ICI 118,551 group. In six dogs, ICI 118,551 was administered as described previously to determine the ability of ₂-adrenoceptor blockade to prevent the increase in ALC produced by terbutaline.

Epinephrine group. These experiments were designed to determine if exogenously administered epinephrine, infused at rates designed to reproduce the pattern of elevated plasma epinephrine concentrations observed after veratrine administration, resulted in a comparable increase in ALC to that observed after veratrine administration. In five dogs, 600 ng/kg/min epinephrine (Sigma) was infused intravenously for 10 min with the infusion rate being decreased to 550 ng/kg/min for the next 50 min. At 1, 2, and 3 h, the infusion rate was further decreased to, respectively, 500, 400, and 360 ng/kg/min. These infusion rates were selected using data of Clutter and coworkers (17) who found that a linear relationship existed between the epinephrine infusion rate and steady-state plasma epinephrine concentration in humans and that at a given infusion rate, plasma epinephrine concentration remained constant once steady-state values had been achieved.

Epinephrine + ICI 118,551 group. In five dogs, ICI 118,551 was administered as described previously to determine the ability of ₂-adrenoceptor blockade to prevent the increase in ALC produced by epinephrine infusion.

Norepinephrine group. These experiments were designed to determine if norepinephrine, infused in doses designed to reproduce the pattern of elevated plasma norepinephrine concentrations observed after veratrine administration, increases ALC. In two dogs, norepinephrine bitartrate (Sigma) was infused intravenously at rates of 151 ng base/kg/min the first hour, 106 ng/kg/min the second hour, and 78 ng/kg/min the remaining 2 h. These infusion rates were selected by constructing a linear regression of plasma norepinephrine concentration on norepinephrine infusion rate using data of Silverberg and coworkers (18).

Statistical Analysis

Potential differences in ALC between groups were evaluated by analysis of variance (ANOVA) followed by a Student-Neuman-Keuls test to determine individual differences where a significant F value was obtained. Plasma catecholamine data were analyzed by repeated measures ANOVA followed by Dunnett's test for multiple comparisons against a single control value. No statistical analyses were done on plasma catecholamine data for experiments in which baseline ALC was evaluated in adrenalectomized animals and those in which norepinephrine was administered, because only two animals were evaluated for each of these interventions. Paired comparisons were made using a paired Student's t test. A p value of less than 0.05 was considered to represent a statistically significant difference.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hemodynamic Changes

Pa started to rise within minutes after veratrine administration. Although the animals were bled from the carotid arterial catheter during this period, extreme increases in Pa developed within 10 min in all of the groups of animals administered veratrine (Figure 1). With the exception of the adrenalectomized group, Pa slowly decreased after this time to baseline levels over the course of the experiment. In the adrenalectomized animals, Pa fell at a faster rate and plateaued at a value that was 36% below baseline pressure. Neither ICI 118,551, amiloride, nor adrenalectomy affected baseline Pa. To prevent the severe pulmonary hypertension that typically develops after veratrine administration (4, 5), the carotid arterial catheter was opened as soon as Ppa started to rise, allowing blood to empty into the reservoir. Figure 2 is a representative pressure tracing showing the ability of the controlled hemorrhage procedure to prevent Ppa and Ppaw from increasing substantially. Figure 1 demonstrates that these pressures were maintained at low levels during the entire experiment. Neither ICI 118,551, amiloride, nor adrenalectomy affected baseline Ppa and Ppaw. Transient increases in cardiac output were observed 10 min after the administration of veratrine in all veratrine groups with the exception of those animals treated with ICI 118,551. No significant changes in cardiac output were observed during this period in any of the control groups.

View larger version (29K): [in this window] [in a new window]   Figure 1.   Vascular pressures and cardiac output before and after veratrine (arrow) administration. Controlled hemorrhage procedure was used to maintain Ppa and Ppaw near baseline values.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Ability of controlled hemorrhage procedure to prevent the development of pulmonary hypertension after veratrine administration. At the arrows marked "O", the carotid arterial catheter stopcock was opened to allow blood to flow into a reservoir. At the arrows marked "C", the stopcock was closed. In this experiment, two hemorrhages were required to stabilize pulmonary arterial pressure. Ppaw = pulmonary arterial wedge pressure determination.

Pa increased an average 28% in the animals in which epinephrine was infused in the presence of ICI 118,551, but did not increase when epinephrine was infused alone (Figure 3). As with the veratrine experiments, it was possible to use the controlled hemorrhage procedure to maintain Ppa and Ppaw close to baseline values (Figure 3). Epinephrine doubled cardiac output, but the increase was prevented by ₂-adrenergic blockade (Figure 3), which resulted in increases in arterial pressure and afterload (19). In the norepinephrine infusion experiments, Pa increased an average 18% 10 min after the start of infusion but returned to baseline values by 90 min (data not shown). Norepinephrine infusion did not alter pulmonary vascular pressures or cardiac output. In experiments in which terbutaline was added to the instilled alveolar fluid, cardiac output was increased an average 15% during the first hour and then gradually returned to baseline values (data not shown). In contrast, cardiac output did not change in experiments in which both ICI 118,551 and terbutaline were administered. The addition of terbutaline to the instilled alveolar fluid, alone or in the presence of ICI 118,551, did not alter systemic or pulmonary vascular pressures.

View larger version (23K): [in this window] [in a new window]   Figure 3.   Effect of epinephrine infusion alone and in the presence of ICI 118,551 on vascular pressures and cardiac output. Controlled hemorrhage procedure was used to maintain Ppa and Ppaw near baseline values.

Alveolar Liquid Clearance

Under baseline conditions in animals anesthetized using pentobarbital sodium (control group, Figure 4), 14.1 ± 2.1 (SE)% of the instilled water volume had been absorbed from the airspaces at the end of the 4-h observation period. (This represented an increase in alveolar fluid protein concentration from 6.63 ± 0.22 to 7.64 ± 0.29 g/dl.) This baseline rate of clearance was similar to that previously observed in both the intact dog (10) and in isolated dog lung lobes (11). Massive SNS activation by the intracisternal injection of veratrine increased the rate of clearance by 116% to 30.4 ± 1.6% (Figure 4). (In these animals, alveolar fluid protein concentration increased from 6.52 ± 0.25 to 9.12 ± 0.32 g/dl. No increase in plasma protein concentration occurred during the experiment under either baseline conditions or after SNS activation.)

View larger version (14K): [in this window] [in a new window]   Figure 4.   ALC in control animals and groups of animals administered veratrine. ADX = adrenalectomized animals. *p < 0.05 compared with control group; **p < 0.05 compared with veratrine group.

The increase in ALC produced by veratrine was inhibited by adrenalectomy and by the administration of the specific ₂-adrenoceptor antagonist ICI 118,551 (Figure 4). Neither of these interventions altered baseline ALC (adrenalectomy: 14.1 ± 2.0%; ICI 118,551: 12.1 ± 2.8%). The addition of amiloride to the plasma instillate also reduced the increase in ALC produced by veratrine (Figure 4) and decreased baseline ALC to 9.7 ± 0.8%. The amiloride-sensitive fractions of the baseline ALC and the increase in ALC produced by veratrine were, respectively, 0.31 and 0.51.

Epinephrine infusion increased ALC (34.7 ± 1.2%) to a similar degree to that which occurred after veratrine administration (Figure 5). The administration of ICI 118,551 markedly reduced the increased ALC produced by epinephrine. Norepinephrine infusion did not increase ALC (9.3 ± 0.8%).

View larger version (10K): [in this window] [in a new window]   Figure 5.   Effect of epinephrine (epi) and epinephrine administered in the presence of ICI 118,551 on ALC. *p < 0.05 compared with control group; **p < 0.05 compared with epinephrine group.

ALC in the control group of animals (n = 8) that was anesthetized with the thiamylal sodium/-chloralose anesthetic combination was 8.1 ± 1.1%. This control group thus formed the basis of comparison for the experiments in which terbutaline was administered because these animals were also anesthetized with this drug combination. Terbutaline, a ₂-adrenergic receptor agonist, also increased ALC (29.9 ± 2.6%, Figure 6), with the increased ALC being reduced by the administration of ICI 118,551. 

View larger version (10K): [in this window] [in a new window]   Figure 6.   Effect of terbutaline and terbutaline administered in the presence of ICI 118,551 on ALC. *p < 0.05 compared with control group; **p < 0.05 compared with terbutaline group.

Comparison of Residual Alveolar Water and Excess Lung Water

The relationship between residual alveolar water (the volume of water in the instilled plasma remaining in the airspaces at the end of the experiment) and excess lung water (the volume of water in the instilled plasma remaining in the lung irrespective of anatomic location) is shown in Figure 7. As indicated by the figure and correlation coefficient (r = 0.92, p < 0.001), there was a strong concordance between the two measures of residual lung water.

View larger version (17K): [in this window] [in a new window]   Figure 7.   Relationship between excess lung water (determined gravimetrically) and residual alveolar water (determined by mass balance); n = 54.

Plasma and Alveolar Fluid Catecholamines

Plasma epinephrine and norepinephrine concentration determinations for each of the experimental groups are shown in Tables 1 and 2, respectively. In the control group, plasma epinephrine concentration remained at baseline values for the entire experiment. Baseline plasma epinephrine concentrations were not altered by adrenalectomy or by the administration of ICI 118,551 or amiloride. Intracisternal veratrine administration produced extreme increases in plasma epinephrine concentration within 10 min of administration that were sustained at high values for the duration of the experiment. SNS activation in the presence of ICI 118,551 or amiloride similarly increased plasma epinephrine concentration, whereas no change was observed after veratrine administration in adrenalectomized dogs. Intravenous epinephrine infusion, either administered alone or in the presence of ICI 118,551, produced a similar pattern of elevated plasma epinephrine concentrations, although the absolute values tended to be higher than those produced by SNS activation. Neither the addition of terbutaline to the instilled alveolar fluid nor the infusion of norepinephrine significantly increased plasma epinephrine concentration.

Veratrine administration also resulted in large increases in plasma norepinephrine concentration that persisted for the duration of the experiment (Table 2). The administration of ICI 118,551 or amiloride did not alter this response, but adrenalectomy attenuated the increase in plasma norepinephrine concentration produced by veratrine. In the two animals in which norepinephrine was infused, plasma norepinephrine concentrations of a similar magnitude to those observed after veratrine administration were produced (Table 2).

In control animals, plasma norepinephrine concentration rose progressively during the experiment, but the magnitude of the increase over the 4-h observation period was substantially less than that observed 10 min after veratrine administration. A similar pattern of progressively increasing plasma norepinephrine concentration was also observed in animals administered epinephrine or terbutaline. The reason for this upward drift is not known but presumably reflects either an increased amount of norepinephrine spillover from sympathetic nerves or a decrease in plasma clearance.

A preliminary analysis of the alveolar fluid epinephrine concentration data indicated that there were no differences between any of the control groups of animals; between the veratrine, veratrine + ICI 118,551, and veratrine + amiloride groups; between the epinephrine and epinephrine + ICI 118,551 groups; and between the terbutaline and terbutaline + ICI 118,551 groups. Accordingly, the data were pooled to form a single control group, veratrine group (with the exception of the adrenalectomized animals), epinephrine group, and terbutaline group and are displayed in Table 3. No increases in alveolar fluid epinephrine concentration were observed after 4 h in the control animals and animals administered either terbutaline or norepinephrine. In contrast, significant increases in alveolar fluid epinephrine concentration were observed at the end of the experiment in the veratrine and epinephrine groups. No increase in epinephrine was observed in adrenalectomized animals administered veratrine.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The major finding of this study is that ALC was increased in a canine model of NPE due to a ₂-adrenoceptor-mediated increase in alveolar epithelial sodium transport, with epinephrine released from the adrenal glands the likely mediator. This conclusion is based on the abilities of adrenalectomy, and ₂-adrenoceptor and sodium channel blockades to inhibit the increased ALC and that of exogenously administered epinephrine to reproduce the response. In the following paragraphs, we discuss this evidence in more detail.

Bilateral adrenalectomy prevented both plasma epinephrine concentration (Table 1) and ALC (Figure 4) from increasing after veratrine administration. Inasmuch as epinephrine is a ₂-adrenergic agonist and has been shown to increase ALC in intact sheep (12) and rats (20) and sodium transport in isolated alveolar epithelial type II cells (8), the results of the adrenalectomy experiments suggest that epinephrine was responsible for increasing ALC after veratrine administration. To further test this hypothesis, we infused epinephrine to produce increases in plasma epinephrine concentration that were of a similar magnitude to those observed after SNS activation (Table 1). In these experiments, ALC was increased to a comparable degree (Figure 5) to that observed after veratrine administration (Figure 4), lending further support to the conclusion that epinephrine released from the adrenal glands mediated the increase in ALC.

Although adrenalectomy prevented plasma epinephrine concentration from increasing after SNS activation, it also significantly reduced the increase in plasma norepinephrine concentration that occurred (Table 2). Accordingly, we also evaluated the possibility that norepinephrine contributed to the increase in ALC by determining fluid clearance in animals in which norepinephrine had been infused to simulate the increased plasma norepinephrine concentrations (Table 2) that occurred after SNS activation in animals with intact adrenals. Norepinephrine infusion did not increase ALC, indicating that circulating norepinephrine did not play a role in increasing ALC after SNS activation. These results are consistent with observations in isolated alveolar epithelial type II cells showing that -adrenergic agonists do not increase alveolar epithelial sodium transport (8).

To determine if the increased ALC was mediated by ₂-adrenoceptors, we tested the ability of the specific ₂-adrenoceptor antagonist ICI 118,551 to prevent ALC from increasing after veratrine administration. In initial experiments, we tested the ability of ICI 118,551 to inhibit the increase in ALC produced by the ₂-adrenergic agonist terbutaline (Figure 6) to verify its ability to block alveolar epithelial ₂-adrenoceptors in the canine lung and found the drug to be an effective antagonist. When administered in the same dose to animals receiving veratrine, ICI 118,551 reduced the increase in ALC (Figure 4), indicating that the increased ALC was mediated by ₂-adrenoceptor activation. Finally, further support for ₂-adrenoceptor involvement was provided by the ability of ICI 118,551 to inhibit the increase in ALC produced by epinephrine infusion (Figure 5).

Amiloride, a sodium channel blocker, was used to determine if the increased ALC produced by SNS activation was mediated by an increase in alveolar epithelial sodium transport. Amiloride reduced the baseline level of liquid clearance 31%, but did not abolish it, suggesting that the canine alveolar epithelium possesses sodium transport processes that are both amiloride-sensitive and insensitive as has been reported for other species (7, 21, 22). In the veratrine-treated animals, amiloride markedly reduced ALC (Figure 4), indicating that the increased clearance was mediated by an increase in sodium transport. The increase in ALC produced by veratrine was inhibited to a greater degree (51%) by amiloride than the baseline level of clearance, suggesting that after massive SNS activation, a greater fraction of the sodium transport traverses the epithelium via amiloride-sensitive pathways compared with that which occurs under baseline conditions. This conclusion is consistent with previous results (7, 21, 22) that suggest that the increase in ALC produced by the -adrenergic agonists isoproterenol and terbutaline is characterized by an increase in the proportion of the total sodium transport that is amiloride-sensitive.

The mechanisms by which -adrenergic agonists increase alveolar epithelial sodium transport have not been completely elucidated, but there is evidence to support the involvement of both apical sodium channels and Na⁺-K⁺-ATPase located in the basolateral membrane of the alveolar epithelial type II cell in mediating the increased sodium flux. In this regard, terbutaline has been found to increase both the open time and open probability of single channels (23) and to increase Na⁺-K⁺-ATPase activity by a sodium-independent mechanism (24) in isolated rat alveolar epithelial type II cells.

What is the fate of the fluid that was absorbed from the airspaces? The results of previous studies suggest that fluid cleared from the alveoli rapidly leaves the lung and is not stored to any appreciable degree in the interstitium (13, 25). This appeared to be the case in our experiments as well. For each experiment, we determined both the residual alveolar lung water (the volume of instilled water remaining in the airspaces at the end of the experiment) and the excess lung water (the volume of instilled water remaining in the lung lobe irrespective of the precise anatomic site) content. Under both baseline conditions and at accelerated rates of clearance, there was a strong concordance between the two measures of lung water (Figure 7), indicating that at the end of the 4-h observation period, essentially none of the absorbed fluid remained in the lung.

As indicated in METHODS, it was critical to prevent alveolar edema from developing in order to determine ALC in experiments in which the SNS was massively activated. This was accomplished by preventing pulmonary vascular pressures from increasing. Immediately preceding the development of NPE, however, severe pulmonary hypertension may develop (1), and recent studies have raised the possibility that extreme increases in pressure might disrupt the integrity of the pulmonary vascular endothelial and alveolar epithelial barriers (26). An increase in epithelial permeability, if it were to occur, may be of some consequence to recovery from edema, because an alveolar epithelium with an abnormally high permeability might have an impaired ability to produce the transepithelial osmotic gradient required to reabsorb excess alveolar fluid. The extent to which epithelial permeability might be altered in NPE, however, is not known. Previous observations indicate that the recovery time course is relatively short in both patients (1) and experimental animals (5) with NPE. This suggests that, in general, any increase in alveolar epithelial permeability that might occur is probably not of a magnitude that would significantly impair the ability of the alveolar epithelium to engage in vectorial sodium and water transport.

We previously observed that the extravascular lung water content, measured by indicator dilution, in dogs with NPE produced by veratrine administration increased markedly within minutes after SNS activation and then appeared to decrease toward baseline levels over the next several hours (5). The results of this study suggest that an increase in alveolar epithelial sodium transport produced by endogenous epinephrine may play a significant role in promoting the rapid recovery. Several lines of evidence suggest that our observations in dogs may be relevant to humans recovering from NPE. First, Sakuma and coworkers (7) have recently found that terbutaline increases ALC in resected human lung lobes, indicating that alveolar epithelial sodium transport is capable of being upregulated by ₂ stimulation in humans. The increase in ALC could be prevented with propranolol, a -adrenergic antagonist. Second, patients with NPE may exhibit elevated plasma and urinary catecholamine concentrations (1). Finally, NPE generally resolves relatively quickly (24 to 72 h) in patients who survive the initial CNS insult (1), suggesting that ALC may be upregulated. Taken together, these observations suggest that the ability to remove sodium and water from the airspaces might be upregulated in patients with NPE.

Our observations indicating that ALC is increased by adrenal epinephrine during recovery from NPE support the emerging concept that the lung possesses intrinsic mechanisms that may enable it not only to maintain an appropriate level of hydration but also to hasten the recovery from pulmonary edema. Other recent observations have begun to provide additional insight into this possibility by identifying forms of pulmonary edema or clinical conditions where the ability of the lung to remove excess alveolar fluid may have been enhanced. In this regard, either ALC or components of the alveolar transepithelial sodium transport process have been found to be upregulated in rat models of sepsis (20), acute bacterial pneumonia (27), hemorrhagic shock (28), alveolar endotoxin administration (29), hyperoxia (30), as well as thiourea (33) and high airway pressure (34) induced pulmonary edema. Our observations in dogs thus indicate ALC can be upregulated endogenously in other species as well. Although epinephrine appears to be responsible for the increase in ALC observed during recovery from NPE in the dog, and endogenous catecholamines have been found to increase ALC in rats with septic (20) and hemorrhagic (28) shock, other mediators may be responsible for modulating ALC in other forms of edema. In this regard, Rezaiguia and coworkers (27) found that tumor necrosis factor-alpha appears to increase ALC in rats with acute bacterial pneumonia.

The observation that the epinephrine concentration of the instilled plasma was increased at the end of the experiments in which either veratrine was administered or epinephrine infused (Table 3) indicated that circulating epinephrine was able to enter the airspaces. Although elevated, the alveolar fluid epinephrine concentration was only 10 to 13% of the plasma epinephrine concentration at the time the alveolar fluid sample was obtained, indicating that the plasma and alveolar fluid epinephrine pools were not equilibrated. The reason for the differences in concentration between the two compartments is not clear. A complicating factor in evaluating possible reasons for the difference is the uncertainty of whether epinephrine entered the airspace fluid via the alveolar capillaries or the bronchial microvasculature.

In conclusion, the results of this study suggest that an adrenal epinephrine-mediated increase in ALC may play an important role in the rapid recovery of NPE observed in the dog (5) and might explain observations from case studies that this form of edema tends to recover quickly (1). Studies showing that it is possible to pharmacologically increase the rate of alveolar fluid reabsorption suggest that it might eventually be possible to devise therapies designed to promote alveolar fluid reabsorption and consequently improve gas exchange and oxygenation in patients with alveolar flooding. In this regard, interest has been expressed in the possible use of ₂-adrenergic agonist therapy in such patients (35). The observation that edema resolution in NPE appears to be naturally accelerated by an endogenous ₂-adrenergic agonist raises questions as to whether it will be possible to make clinically useful improvements in the rate of recovery from alveolar flooding in patients with NPE or other forms of edema in which ALC has been found to be endogenously accelerated.