Vascular and extravascular compartments of the isolated perfused rabbit lung ROBERT S. Y. CHANG, PHILLIP SILVERMAN, AND RICHARD M. EFFROS Division of Respiratory Physiology and Medicine, Department of Medicine, Harbor General Hospital-UCLA School of Medicine, Torrance, California 90509 CHANG, ROBERT S. Y., PHILLIP SILVERMAN, AND RICHARD M. EFFROS. Vascular and extravascular compartments of the isolated perfused rabbit lung. J. Appl. Physiol.: Respirat. Environ. Exercise Physiol. 46(l): 74-78, 1979. -The vascular, interstitial, and cellular compartments of 15 isolated, perfused and ventilated rabbit lungs were determined by a steady-state indicator-dilution procedure. Five lungs were perfused with constant pulmonary artery flow and zero left atria1 pressure for more than 1 h. Edema formation was continuous and pulmonary vascular volume (PVV) decreased initially at a time when pulmonary vascular resistance (PVR) was falling. Increases in PVR were not seen until edema formation had become severe. In 10 other lungs, increases in pulmonary artery or left atria1 pressure resulted in elevation of PVV and accelerated edema formation. The initially abrupt increase in PVV was followed by a more gradual increase over a lo-min period. Return of fluid to the vasculature was never observed in these studies. Labeled albumin readily entered the extravascular space but a relatively constant fraction of the interstitium remained inaccessible to albumin. No changes were found in the cellular volume during edema formation. pulmonary edema; indicator vascular volume; pulmonary
dilution, steady-state; circulation
pulmonary
ISOLATED AND PERFUSED LUNG preparation has assumed an increasingly important role in studies of pulmonary hemodynamics, metabolism, and pharmacology (18). The popularity of this experimental approach is understandable: pulmonary parameters can be studied apart from the effects of other organs and the conditions to which the lung is exposed can be altered with great precision. One major disadvantage of isolated and perfused lungs is the nearly universal observation that they tend to become edematous after a variable period of time. Edema formation may be related to lymphatic dysfunction (23), the character of the perfusion fluid (12), the mode of ventilation (1, 21), or any number of other factors that are not well understood. It is common practice to weigh the lungs during a study to detect fluid accumulation (2, 5, 7, 11). As emphasized by Staub (16) this procedure is fraught with uncertainty because there is no way to discriminate, on the basis of weight alone, between changes in the vascular, interstitial, or cellular volumes. It was the objective of this study to develop a method that would permit discrete measure-
THE
74
ments of changes in these volumes under a variety of circumstances. We have used this procedure to study the response of the rabbit lung to variations in pulmonary arterial and venous hydrostatic pressures. MATERIALS
AND
METHODS
Fifteen New Zealand albino rabbits weighing between 2 and 3.6 kg were anesthetized with pentobarbital (lo-20 mg/kg) administered through a marginal ear vein with additional doses given as needed. A tracheostomy was performed and the animals were ventilated with an animal respiratory pump (Harvard, Millis, MA) using 10 cmH,O peak inspiratory pressure and 3 cmH,O end-expiratory pressure. Porcine heparin, 4,000 U in 2 ml saline, was administered intravenously and the animals were exsanguinated via a femoral artery catheter. A midline thoracotomy was immediately performed and cannulas were secured in the pulmonary artery and left atrium to serve as the inflow and venous return, respectively. The pulmonary artery was temporarily constricted with suture material placed around the vessel; the right ventricular outflow was flushed thoroughly before inserting the pulmonary artery cannula to avoid the introduction of air emboli into the circulation (18). The lung was then washed out with 100 ml of a physiological albumin solution, excised, and placed in a Plexiglas chamber heated to 37°C (see Fig. 1). The perfusion fluid was made of equal quantities of plasma obtained from the rabbit and an albumin solution which contained: 5.8 g/l NaCl, 0.30 g/l KCl, 2.5 g/l NaHCOz, 37 g/l CaCl,, 20 g/l MgS04, 1.5 g/l glucose, 50 g/l bovine albumin (Cohn Fraction V, Sigma Chemical Co., St. Louis, MO), 5,000 IU/l porcine heparin. The quantity of perfusion fluid was kept as small as possible (66 t 9 (SD) ml) to increase the sensitivity of measurements of indicator dilution. A constant-flow pump (Cole-Parmer peristaltic pump, Chicago, IL) was used to minimize perfusate volume and to permit direct calculations of total pulmonary vascular resistance from pulmonary artery pressure data. In a manner similar to Weiser and Grande (22), we chose to weigh the perfusate reservoir rather than the lung. This procedure served to eliminate errors related to the tendency for the arterial and venous lines to support the lung weight. A small T tube was placed in the arterial line and covered with a soft rubber mem-
0161-7567/79/0000-OOOOooo$Ol.25 Copyright
0 1979 the American
Physiological
Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 11, 2019.
FLUID
COMPARTMENTS
IN ISOLATED
PERFUSED
RABBIT
LUNGS
brane. This dampened oscillations of arterial pressure and also kept air bubbles from reaching the pulmonary artery. Perfusion was begun 10 min before the addition of indicators at a constant pulmonary artery flow. Reservoir weight, pulmonary artery and left atria1 pressures, pump speed, and tracheal pressures were monitored continuously during the course of the study and were displayed on a polygraph (Beckman R-611, Schiller Park, IL). At the onset of the study period, a mixture of indicators was introduced as a bolus of 20 ml into the reservoir. This injection solution contained a 1:l mixture of autologous rabbit plasma and the 5 g/d1 albumin solution. In addition, the following indicators were included: dextran blue (Pharmacia, Sweden; mol wt 2 million, 1 .O g/l), 22Na (New England Nuclear, Boston, MA; 47 &iI 1) and 12”1-albumin (New England Nuclear; 0.1 &i/l). In the studies of pulmonary artery pressure changes, two additional indicators were used: [N-methyl 14C]antipyrine (New England Nuclear; 30 &i/l) and ggmTc (New England Nuclear; 25 &i/l) chelated with diethylenetriaminepentaacetate (DTPA) (Union Carbide Corp., Tuxedo, NY). Dextran blue has been used as a vascular indicator in transient indicator-dilution studies of the lung (13). Although red blood cells have been used to measure the vascular space (11, 20) our experience with ggmTc- or “Cr-labeled red blood cells has been unsatisfactory in this preparation. In each of five preliminary studies, the apparent volume of distribution of red blood cells significantly exceeded that of 12”1-albumin. This could be due either to trapping of red blood cells in the vasculature, or loss of the label into the tissues secondary to hemolysis or dissociation of label from the red blood cells. Furthermore, we would prefer to avoid using red blood cells because these may alter studies of pulmonary metabolism and pharmacology (18). It must be stressed that dextran blue, or even red blood cells, may leak from the pulmonary vasculature, resulting in
4 ,,
I@&&
-TACHOMETER PERISTALTIC
PyP
II
37OC
v FIG.
See text reservoir,
PERISTALTIC
BOX
PUMP
1. Isolated lungs are placed in a heated Plexiglas chamber. for detailed description. Note that weight of perfusate instead of the lung, is monitored during these studies.
75
overestimates of the pulmonary vascular volume and underestimates of filtration rates. Samples obtained from the reservoir were spun at 1,000 x g for 20 min. The supernatant fractions were read at 630 nm in a spectrophotometer (Bausch and Lomb Spectronic 700 with digital output, Rochester, NY). The optical density (OD) of the perfusion fluid in the absence of the blue dye was between 0.02 and 0.04 in these experiments. Following the addition of dye, ODs increased to between 0.300 and 0.800. The background color was due in part to hemolysis of the small number of remaining red blood cells in the preparation (hemoglobin co.05 g/dl). In three separate studies, in which no dextran blue was added, the OD of the perfusion fluid was measured to determine whether changes in the “background” occurred over a period of 1 h. These measurements did not vary more than 0.004 OD Bovine albumin, labeled with 1251, was used as an indicator for the volume of distribution of albumin. The 1251label was 96% precipitable with 10% trichloroacetic acid. 22Na+ and, in some studies, ggmTc-DTPA were used to measure the extracellular volume. ggmTc-DTPA has a molecular weight of492 and is cleared by the kidneys in a manner similar to other extracellular indicators such as “Cr-EDTA and inulin (9, 15,19). To provide an index of the total water volume of the preparation, [14C]antipyrine was used. We found that during the first 10 min of perfusion, extravascular volume of distribution of this indicator is 7% greater than that of 3H20. This is similar to the relative distributions of antipyrine and 3H20 in red blood cells (4) and may be due to protein binding (3). Our decision to use [14C]antipyrine was based on the observation that “H,O exchanges with water vapor in the inspired air during the course of the studies and, therefore, could not be used to measure the water content of the lung. The activities of the gamma isotopes were determined in an automated three-channel gamma scintillation counter (Intertechnique, Paris, France). The beta activities were measured in an automated liquid scintillation counter (Packard, Downer’s Grove, IL). Crossover corrections were made with the aid of a Hewlett-Packard 3000 computer. Standards were made by diluting the injection solution in fluid which was removed from the reservoir 5 min before addition of indicators. The observation that the 12”1-albumin volume was slightly smaller than the dextran blue volume at the beginning of some studies was attributable to an excessive amount of hemoglobin in the fluid used to dilute the standards in these studies. Aliquots of 0.8 ml were removed from the reservoir at 5min intervals. The volume of dilution (Vi) of an indicator (i) was calculated with the equation: Vi(t) = Mi( t>/Ci( t), where Mi( t) is the amount of indicator in the preparation at time t divided by the fractional concentration observed at that time, Ci( t). Mi( t) is corrected for the amount of indicator removed by sampling prior to time t. The indicator-dilution volumes of each indicator in the extravascular space, VEVi( t), are calculated from the total
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 11, 2019.
76
CHANG,
indicator volumes by subtracting umes. For example vEVzzNa+
(
t)
=
vqa+
the dextran (
t)
-
VDB(
SILVERMAN,
AND
EFFROS
blue vol-
t)
The pulmonary vascular volume, PVV(t), was calculated by subtracting the constant volume of fluid in the tubing, Vd, and the reservoir volume, Vr( t), from the total perfusate volume derived from dextran blue dilution PVV(t)
= V,,(t)
- Vr(t)
- Vd
Vr( t) was calculated by dividing the weight of the fluid in the reservoir by the specific gravity of the perfusion solution. The rate of filtration of fluid into the extravascular space was calculated from the decline of perfusate volume, VDB, with time. The coefficients of variation of indicator concentrations in our standard dilutions of the injection solution averaged 0.011, 0.018, 0.031, 0.011, 0.016 for dextran blue, 1251-albumin, 22Na, ggmTc-DTPA, and [ 14C]antipyrine, respectively. Three groups of isolated rabbit lungs were studied: a) control lungs with constant pulmonary artery flow and zero left atria1 pressure; b) lungs in which pulmonary artery flow was periodically elevated to produce increases in pulmonary artery pressure; and c) lungs perfused at constant flow with periodic elevations of left atria1 pressure. RESULTS
Constant-flow studies (Fig. 2). Pulmonary artery flow was kept constant at an average value of 86 -)- 18 of ml/min. During the first 30 min after introduction
~~~$~%-*~~--~
*--*,IrrII*~$~$$g
f
3 $
s~~~~~*$~~flI~
TIME
(MINI
2. Constant-flow studies. Continuous filtration (FR) and edema formation occurred ~during these studies. Decreases in pulmonary artery pressure (Ppa) and pulmonary vascular volume (PVV) were significant by paired t test (see text). Although extravascular volumes of 22Na+ and 1251-albumin increased progressively, difference between 22Na+ and 1251-albumin volumes remained unchanged. Mean values + SE are shown FIG.
TIME
(MIN)
3. Pulmonary artery pressure studies. Only greater inin Ppa lead to significant increases in PVV and filtration rate. Extravascular volume (EVV) of 22Na+, ggmTc-DTPA, [14C]antipyrine, and 1251-albumin increase throughout the studies. A slight increase in the difference between 22Na+ and 1251-albumin volumes occurred in these studies (P < 0.05). FIG.
Icreases
indicators into the reservoir, there was a fall in the pulmonary artery pressure, indicating a decline in the pulmonary vascular resistance (PVR). There was a concomitant decline in the PW, which became significant by 15 min (P < 0.05, paired Student’s t test). Throughout the perfusion period there was an increase in the extravascular volumes of distribution accessible to 1251-albumin and 22Na +. Filtration persisted during these studies but was most rapid at the beginning of the study (P < 0.05). The difference between the 22Na+ and 1251-albumin volumes remained unchanged. At the end of the study, edema fluid was found in the trachea and bronchi of all of these lungs and the wet weight-to-dry weight ratio was 5.22 t 0.49. This compares with control values obtained in five additional studies: wet weight-to-dry weight ratios in these blood-free preparations, which were perfused for 5 min after the addition of indicators, averaged 4.97 t 0.11. The ratio of extravascular sodium to extravascular water was 0.30 t 0.03 in these control lungs (estimated from the quantity of 1251-albumin, 22Na+, and water retained in the drained 111nfA IUU~/.
Pulmonary arterial pressure studies (Fig. 3). The initial pulmonary artery flows in these five lungs aver-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 11, 2019.
FLUID
COMPARTMENTS
IN ISOLATED
PERFUSED
RABBIT
aged 65 ml/min and the pulmonary arterial pressure (Ppa) measurements were 10 t 2 Torr. Small increases in Ppa (6 t 1 Torr) caused an apparent increase in mean PVV and filtration rate but these changes were not significant. Larger increases in Ppa (10 t 3 Torr) resulted in increases in PVV and filtration rate. Although most of the increase in PVV occurred during the first 5 min of each study period, PVV continued to rise during the subsequent 5 min in 9 of the 10 periods in which Ppa was elevated. Similarly slow declines in PVV were observed after flow was returned to control values. The extravascular volumes of distribution of lz51-albumin ggmTc-DTPA, 22Na+, and [ 14C]antipyrine increased throughout the perfusion period. There was a slight increase in the extravascular volumes of the lung which were accessible to ggmTc-DTPA, and 22Na+ but not The cellular volume (estiaccessible to 12”1-albumin. mated from the difference between the [14C]antipyrine volume and the interstitial volumes) remained unchanged. The wet weight-to-dry weight ratios of these lungs averaged 5.46 t 0.44. Left atria1 pressure studies (Fig. 4). Flows in these five studies averaged 63 ml/min. Moderate increases of left atria1 pressure (Pla) (6 * 0.5 Torr) had no signifi-
* ,
JO -
'
T
1T
.
T
PL
10 -
TIME (MIN\ 4. Left atria1 pressure studies. Increases in left atria1 pressure obtained during second period of pressure elevation resulted in statistically significant increases in Ppa, PVV, and filtration rate. control (FR, asterisk indicates significant difference from previous period.) EVV of 22Na+ and 1251-albumin increase in a manner similar to constant-flow and Ppa studies. FIG.
LUNGS
77
cant effect on Ppa or filtration. Greater increases in Pla (13 t 4 Torr) produced increases in Ppa, PVV, and the filtration rate. These parameters returned to normal values during the final control interval. Increases in the extravascular volumes of 1251-albumin and 22Na+ resembled those found in the isoflow and pulmonary artery studies. The wet weight-to-dry weight ratio averaged 5.83 t 0.97. DISCUSSION
The most striking characteristic of this lung preparation was the slow but progressive edema formation observed in all studies. As indicated above, the cause for edema formation in this and other isolated lungs remains uncertain. The constant-flow studies suggest that previous measurements of filtration, based on weight gain alone, are subject to error. We observed an initial decline in PVV that would diminish increases in lung weight related to passage of fluid into the tissues. Edema formation might, therefore, be underestimated by lung weight measurements. Furthermore, the increase in wet weight-to-dry weight ratios was very modest, when compared to controls, despite the presence of gross tracheal edema. This may also reflect the fact that the vasculature is diminished in drained, edematous lungs. It remains to be determined whether vascular contraction with edema formation occurs in isolated lungs of other species, or under other conditions. The decline of PVV occurred despite a fall in overall PVR. It would appear that early decreases in PVV are not accompanied by increases in PVR in this preparation. Although the PVV decline may reflect vascular compression by edema fluid, PVR failed to increase until gross edema formation was observed, and did not seem to be a sensitive parameter for the detection of transvascular filtration. The failure of PVV to decline further may be due, in part, to leakage of dextran blue into the tissues in the later stages of the study. Elevations of Pla and Ppa appear to produce a rapid expansion of the PVV followed by a more gradual increase in PVV that continued for as long as 15 min. Evidence has been obtained that the distension or recruitment of pulmonary vessels with increased hydrostatic pressure may occur over a relatively prolonged period of time (14). Although it has been proposed that pulmonary vascular filling is complete during the first 2 min (7, 8, 11), our study suggests that this may not be the case, and that the slow increase in lung weight used by others to detect filtration may include increases in the PVV. Modest elevations (6 Torr) in Pla produced small increases in mean Ppa, PVV, and filtration rate but none of these changes were significant. Greater elevations in Pla (13 Torr) produced significant increases in each of these parameters. Similarly, only large increases (10 Torr) in Ppa produced measurable increases in PVV and filtration rate. The decrease observed in the rate of filtration in the constant-flow studies may be due to a rise in interstitial pressure as edema accumulates. Alternatively, the per-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 11, 2019.
78
CHANG,
fused capillary surface area may diminish as edema becomes more severe. An apparent decline in filtration might also represent leakage of dextran blue into the interstitium in the later stages of edema formation. The extravascular volumes of distribution available to 22Na+ ggmTc-DTPA, and 1251-albumin increase progressively throughout these studies. The concentrations of each of these solutes in the edema fluid were close to those in the perfusion fluid. No change was found in the cellular volume, indicating that edema formation involved only the interstitial and/or alveolar compartments. Although the capillaries appeared to be permeable to solutes as large as albumin, a relatively constant portion of the extravascular space remained inaccessible to ggmTc-DTPA and (to an even greater extent) to 1251-albumin. It has been suggested that the interstitial matrix excludes larger molecules in a manner analo-
SILVERMAN,
AND
EFFROS
gous to Sephadex particles (10). The integrity of this volume appears to remain intact despite significant edema formation. Evidence has been provided that edema fluid may selectively enter the perivascular and peribronchiolar spaces (17), and that filling of the alveoli may occur after the bronchiolar wall has been breached (6). Bulk movement of protein and fluid may be largely confined to these channels, leaving the interstitial gel relatively unaltered. This research was supported by National Institutes of Health Grants HL-18606 and HL-00132 and by Grant GLAA562 from the A merican Heart Association, Greater Los Angeles Affiliate. We gratefully acknowledge the technical assistance of Ms. Kathy Wright and the fine secretarial work of Mrs. Abbey Rille. Received
28 February
1978; accepted
in final form 21 August
1978.
REFERENCES 1. Bij, G., A. HAUGE, AND G. NICOLAYSEN. Alveolar pressure and lung volume as determinants of net transvascular fluid filtration. J. Appl. Physiol.: Respirat. Environ. Exercise Physiol. 42: 476-482, 1977. 2. BORN, G. V. R. Acute edema in the isolated perfused lungs of rabbits. J. Physiol. London 124: 502-514, 1954. 3. BRODIE, B. B., E. Y. BEYER, J. ANEBROD, M. F. DUNNINE, Y. POROWSKA, AND J. M. STEELE. Use of n-acetyl 4-aminopyrine (NAAP) in measurement of total body water. Proc. Sot. Exp. Biol. Med. 77: 794-798, 1951. 4. EFFROS, R. M., AND F. P. CHINARD. The in vivo pH of the extravascular space of the lung. J. CZin. Invest. 48: 1983-1996, 1969. 5. GAAR, K. A., A. E. TAYLOR, L. J. OWENS, AND A. C. GUYTON. Pulmonary capillary pressure and filtration coefficient in the isolated perfused lung. Am. J. Physiol. 213: 910-914, 1967. 6. GEE, M. H., AND N. C. STAUB. Role of bulk fluid flow in protein permeability of the dog lung alveolar membrane. J. AppZ. Physiol.: Respirat. Environ. Exercise Physiol. 42: 144-149, 1977. 7. GOLDBERG, H. S., W. MITZNER, AND G. BATRA. Effect of transpulmonary and vascular pressures on rate of pulmonary edema formation. J. Appl. Physiol.: Respirat. Environ. Exercise PhysioZ. 43: 14-19, 1977. 8. HAUGE, A., G. Bij, AND B. A. WAALER. Interrelations between pulmonary liquid volumes and lung compliance. J. AppZ. PhysioZ. 38: 608-614, 1975. 9. KLOPPER, J. F., W. HAUSER, H. L. ATKINS, W. C. ECKELMAN, AND P. RICHARDS. Evaluation of ggmTc-DTPA for the measurement of glomerular filtration rate. J. NucZ. Med. 13: 107-110, 1972. 10. LAURENT, T. The structure and function of extracellular polysaccharides in connective tissue. In: CapiZZary Permeability, edited by C. Crone and N. A. Lassen. New York: Academic, 1970, p. 261-277. 11. LUNDE, P. K. M., AND B. A. WAALER. Transvascular fluid balance in the lung. J. Physiol. London 205: l-18, 1969. 12. NICOLAYSEN, G. Intravascular concentrations of calcium and magnesium ions and edema formation in isolated lungs. Acta
PhysioZ. Stand. 81: 325-339, 1971. 13. RYAN, J. W., U. SMITH, AND R. S. NIEMEYER. Angiotensin I: metabolism by plasma membrane of lung. Science 176: 64-66, 1972. 14. SARNOFF, S. J., AND E. BERGLUND. Pressure-volume characteristics and stress relaxation in the pulmonary vascular bed of the dog. Am. J. PhysioZ. 171: 238-244, 1952. 15. STACY, B. D., AND G. D. THOBURN. Chromium-51 ethylenediaminetetraacetate for estimation of glomerular filtration rate. Science 152: 1076-1077, 1966. 16. STAUB, N. C. Pulmonary edema. PhysioZ. Rev. 54: 678-811, 1974. 17. STAUB, N. C., H. NAGANO, AND M. L. PEARCE. Pulmonary edema in dogs, especially the sequence of fluid accumulation in the lungs. J. AppZ. PhysioZ. 22: 227-240, 1967. 18. TIERNEY, D. F., S. L. YOUNG, J. J. O’NEIL, AND M. ABE. Isolated perfused lung-substrate utilization. Federation Proc. 36: 161165, 1977. 19. TRAP-JENSEN, J., AND N. A. LASSEN. Capillary permeability for smaller hydrophilic tracers in exercising skeletal muscle in normal man and in patients with long-term diabetes mellitus. In: CapiZZary Permeability, edited by C. Crone and N. A. Lassen. New York: Academic, 1970, p. 135-152. 20. WAALER, B. A., AND P. AARSETH. Interstitial fluid and transcapillary fluid balance. In: Lung Liquids. Ciba Foundation Symposium, edited by C. J. Dickinson. London: Elsevier, 1975, p. 6571. 21. WEBB, H. H., AND D. F. TIERNEY. Experimental pulmonary edema due to intermittent positive pressure ventilation with high inflation pressures. Protection by positive end-expiratory pressure. Am. Rev. Respir. Dis. 110: 556-565, 1974. 22. WEISER, P. C., AND F. GRANDE. Estimation of fluid shifts and protein permeability during pulmonary edemagenesis. Am. J. PhysioZ. 226: 1028-1034, 1974. 23. WOOLVERTON, W., K. L. BRIGHAM, AND N. C. STAUB. Effect of continuous positive airway pressure breathing (CPAPB) on pulmonary fluid filtration and content in sheep (Abstract). PhysioZogist 16: 490, 1973.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 11, 2019.
dilution, steady-state; circulation
pulmonary
ISOLATED AND PERFUSED LUNG preparation has assumed an increasingly important role in studies of pulmonary hemodynamics, metabolism, and pharmacology (18). The popularity of this experimental approach is understandable: pulmonary parameters can be studied apart from the effects of other organs and the conditions to which the lung is exposed can be altered with great precision. One major disadvantage of isolated and perfused lungs is the nearly universal observation that they tend to become edematous after a variable period of time. Edema formation may be related to lymphatic dysfunction (23), the character of the perfusion fluid (12), the mode of ventilation (1, 21), or any number of other factors that are not well understood. It is common practice to weigh the lungs during a study to detect fluid accumulation (2, 5, 7, 11). As emphasized by Staub (16) this procedure is fraught with uncertainty because there is no way to discriminate, on the basis of weight alone, between changes in the vascular, interstitial, or cellular volumes. It was the objective of this study to develop a method that would permit discrete measure-
THE
74
ments of changes in these volumes under a variety of circumstances. We have used this procedure to study the response of the rabbit lung to variations in pulmonary arterial and venous hydrostatic pressures. MATERIALS
AND
METHODS
Fifteen New Zealand albino rabbits weighing between 2 and 3.6 kg were anesthetized with pentobarbital (lo-20 mg/kg) administered through a marginal ear vein with additional doses given as needed. A tracheostomy was performed and the animals were ventilated with an animal respiratory pump (Harvard, Millis, MA) using 10 cmH,O peak inspiratory pressure and 3 cmH,O end-expiratory pressure. Porcine heparin, 4,000 U in 2 ml saline, was administered intravenously and the animals were exsanguinated via a femoral artery catheter. A midline thoracotomy was immediately performed and cannulas were secured in the pulmonary artery and left atrium to serve as the inflow and venous return, respectively. The pulmonary artery was temporarily constricted with suture material placed around the vessel; the right ventricular outflow was flushed thoroughly before inserting the pulmonary artery cannula to avoid the introduction of air emboli into the circulation (18). The lung was then washed out with 100 ml of a physiological albumin solution, excised, and placed in a Plexiglas chamber heated to 37°C (see Fig. 1). The perfusion fluid was made of equal quantities of plasma obtained from the rabbit and an albumin solution which contained: 5.8 g/l NaCl, 0.30 g/l KCl, 2.5 g/l NaHCOz, 37 g/l CaCl,, 20 g/l MgS04, 1.5 g/l glucose, 50 g/l bovine albumin (Cohn Fraction V, Sigma Chemical Co., St. Louis, MO), 5,000 IU/l porcine heparin. The quantity of perfusion fluid was kept as small as possible (66 t 9 (SD) ml) to increase the sensitivity of measurements of indicator dilution. A constant-flow pump (Cole-Parmer peristaltic pump, Chicago, IL) was used to minimize perfusate volume and to permit direct calculations of total pulmonary vascular resistance from pulmonary artery pressure data. In a manner similar to Weiser and Grande (22), we chose to weigh the perfusate reservoir rather than the lung. This procedure served to eliminate errors related to the tendency for the arterial and venous lines to support the lung weight. A small T tube was placed in the arterial line and covered with a soft rubber mem-
0161-7567/79/0000-OOOOooo$Ol.25 Copyright
0 1979 the American
Physiological
Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 11, 2019.
FLUID
COMPARTMENTS
IN ISOLATED
PERFUSED
RABBIT
LUNGS
brane. This dampened oscillations of arterial pressure and also kept air bubbles from reaching the pulmonary artery. Perfusion was begun 10 min before the addition of indicators at a constant pulmonary artery flow. Reservoir weight, pulmonary artery and left atria1 pressures, pump speed, and tracheal pressures were monitored continuously during the course of the study and were displayed on a polygraph (Beckman R-611, Schiller Park, IL). At the onset of the study period, a mixture of indicators was introduced as a bolus of 20 ml into the reservoir. This injection solution contained a 1:l mixture of autologous rabbit plasma and the 5 g/d1 albumin solution. In addition, the following indicators were included: dextran blue (Pharmacia, Sweden; mol wt 2 million, 1 .O g/l), 22Na (New England Nuclear, Boston, MA; 47 &iI 1) and 12”1-albumin (New England Nuclear; 0.1 &i/l). In the studies of pulmonary artery pressure changes, two additional indicators were used: [N-methyl 14C]antipyrine (New England Nuclear; 30 &i/l) and ggmTc (New England Nuclear; 25 &i/l) chelated with diethylenetriaminepentaacetate (DTPA) (Union Carbide Corp., Tuxedo, NY). Dextran blue has been used as a vascular indicator in transient indicator-dilution studies of the lung (13). Although red blood cells have been used to measure the vascular space (11, 20) our experience with ggmTc- or “Cr-labeled red blood cells has been unsatisfactory in this preparation. In each of five preliminary studies, the apparent volume of distribution of red blood cells significantly exceeded that of 12”1-albumin. This could be due either to trapping of red blood cells in the vasculature, or loss of the label into the tissues secondary to hemolysis or dissociation of label from the red blood cells. Furthermore, we would prefer to avoid using red blood cells because these may alter studies of pulmonary metabolism and pharmacology (18). It must be stressed that dextran blue, or even red blood cells, may leak from the pulmonary vasculature, resulting in
4 ,,
I@&&
-TACHOMETER PERISTALTIC
PyP
II
37OC
v FIG.
See text reservoir,
PERISTALTIC
BOX
PUMP
1. Isolated lungs are placed in a heated Plexiglas chamber. for detailed description. Note that weight of perfusate instead of the lung, is monitored during these studies.
75
overestimates of the pulmonary vascular volume and underestimates of filtration rates. Samples obtained from the reservoir were spun at 1,000 x g for 20 min. The supernatant fractions were read at 630 nm in a spectrophotometer (Bausch and Lomb Spectronic 700 with digital output, Rochester, NY). The optical density (OD) of the perfusion fluid in the absence of the blue dye was between 0.02 and 0.04 in these experiments. Following the addition of dye, ODs increased to between 0.300 and 0.800. The background color was due in part to hemolysis of the small number of remaining red blood cells in the preparation (hemoglobin co.05 g/dl). In three separate studies, in which no dextran blue was added, the OD of the perfusion fluid was measured to determine whether changes in the “background” occurred over a period of 1 h. These measurements did not vary more than 0.004 OD Bovine albumin, labeled with 1251, was used as an indicator for the volume of distribution of albumin. The 1251label was 96% precipitable with 10% trichloroacetic acid. 22Na+ and, in some studies, ggmTc-DTPA were used to measure the extracellular volume. ggmTc-DTPA has a molecular weight of492 and is cleared by the kidneys in a manner similar to other extracellular indicators such as “Cr-EDTA and inulin (9, 15,19). To provide an index of the total water volume of the preparation, [14C]antipyrine was used. We found that during the first 10 min of perfusion, extravascular volume of distribution of this indicator is 7% greater than that of 3H20. This is similar to the relative distributions of antipyrine and 3H20 in red blood cells (4) and may be due to protein binding (3). Our decision to use [14C]antipyrine was based on the observation that “H,O exchanges with water vapor in the inspired air during the course of the studies and, therefore, could not be used to measure the water content of the lung. The activities of the gamma isotopes were determined in an automated three-channel gamma scintillation counter (Intertechnique, Paris, France). The beta activities were measured in an automated liquid scintillation counter (Packard, Downer’s Grove, IL). Crossover corrections were made with the aid of a Hewlett-Packard 3000 computer. Standards were made by diluting the injection solution in fluid which was removed from the reservoir 5 min before addition of indicators. The observation that the 12”1-albumin volume was slightly smaller than the dextran blue volume at the beginning of some studies was attributable to an excessive amount of hemoglobin in the fluid used to dilute the standards in these studies. Aliquots of 0.8 ml were removed from the reservoir at 5min intervals. The volume of dilution (Vi) of an indicator (i) was calculated with the equation: Vi(t) = Mi( t>/Ci( t), where Mi( t) is the amount of indicator in the preparation at time t divided by the fractional concentration observed at that time, Ci( t). Mi( t) is corrected for the amount of indicator removed by sampling prior to time t. The indicator-dilution volumes of each indicator in the extravascular space, VEVi( t), are calculated from the total
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 11, 2019.
76
CHANG,
indicator volumes by subtracting umes. For example vEVzzNa+
(
t)
=
vqa+
the dextran (
t)
-
VDB(
SILVERMAN,
AND
EFFROS
blue vol-
t)
The pulmonary vascular volume, PVV(t), was calculated by subtracting the constant volume of fluid in the tubing, Vd, and the reservoir volume, Vr( t), from the total perfusate volume derived from dextran blue dilution PVV(t)
= V,,(t)
- Vr(t)
- Vd
Vr( t) was calculated by dividing the weight of the fluid in the reservoir by the specific gravity of the perfusion solution. The rate of filtration of fluid into the extravascular space was calculated from the decline of perfusate volume, VDB, with time. The coefficients of variation of indicator concentrations in our standard dilutions of the injection solution averaged 0.011, 0.018, 0.031, 0.011, 0.016 for dextran blue, 1251-albumin, 22Na, ggmTc-DTPA, and [ 14C]antipyrine, respectively. Three groups of isolated rabbit lungs were studied: a) control lungs with constant pulmonary artery flow and zero left atria1 pressure; b) lungs in which pulmonary artery flow was periodically elevated to produce increases in pulmonary artery pressure; and c) lungs perfused at constant flow with periodic elevations of left atria1 pressure. RESULTS
Constant-flow studies (Fig. 2). Pulmonary artery flow was kept constant at an average value of 86 -)- 18 of ml/min. During the first 30 min after introduction
~~~$~%-*~~--~
*--*,IrrII*~$~$$g
f
3 $
s~~~~~*$~~flI~
TIME
(MINI
2. Constant-flow studies. Continuous filtration (FR) and edema formation occurred ~during these studies. Decreases in pulmonary artery pressure (Ppa) and pulmonary vascular volume (PVV) were significant by paired t test (see text). Although extravascular volumes of 22Na+ and 1251-albumin increased progressively, difference between 22Na+ and 1251-albumin volumes remained unchanged. Mean values + SE are shown FIG.
TIME
(MIN)
3. Pulmonary artery pressure studies. Only greater inin Ppa lead to significant increases in PVV and filtration rate. Extravascular volume (EVV) of 22Na+, ggmTc-DTPA, [14C]antipyrine, and 1251-albumin increase throughout the studies. A slight increase in the difference between 22Na+ and 1251-albumin volumes occurred in these studies (P < 0.05). FIG.
Icreases
indicators into the reservoir, there was a fall in the pulmonary artery pressure, indicating a decline in the pulmonary vascular resistance (PVR). There was a concomitant decline in the PW, which became significant by 15 min (P < 0.05, paired Student’s t test). Throughout the perfusion period there was an increase in the extravascular volumes of distribution accessible to 1251-albumin and 22Na +. Filtration persisted during these studies but was most rapid at the beginning of the study (P < 0.05). The difference between the 22Na+ and 1251-albumin volumes remained unchanged. At the end of the study, edema fluid was found in the trachea and bronchi of all of these lungs and the wet weight-to-dry weight ratio was 5.22 t 0.49. This compares with control values obtained in five additional studies: wet weight-to-dry weight ratios in these blood-free preparations, which were perfused for 5 min after the addition of indicators, averaged 4.97 t 0.11. The ratio of extravascular sodium to extravascular water was 0.30 t 0.03 in these control lungs (estimated from the quantity of 1251-albumin, 22Na+, and water retained in the drained 111nfA IUU~/.
Pulmonary arterial pressure studies (Fig. 3). The initial pulmonary artery flows in these five lungs aver-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 11, 2019.
FLUID
COMPARTMENTS
IN ISOLATED
PERFUSED
RABBIT
aged 65 ml/min and the pulmonary arterial pressure (Ppa) measurements were 10 t 2 Torr. Small increases in Ppa (6 t 1 Torr) caused an apparent increase in mean PVV and filtration rate but these changes were not significant. Larger increases in Ppa (10 t 3 Torr) resulted in increases in PVV and filtration rate. Although most of the increase in PVV occurred during the first 5 min of each study period, PVV continued to rise during the subsequent 5 min in 9 of the 10 periods in which Ppa was elevated. Similarly slow declines in PVV were observed after flow was returned to control values. The extravascular volumes of distribution of lz51-albumin ggmTc-DTPA, 22Na+, and [ 14C]antipyrine increased throughout the perfusion period. There was a slight increase in the extravascular volumes of the lung which were accessible to ggmTc-DTPA, and 22Na+ but not The cellular volume (estiaccessible to 12”1-albumin. mated from the difference between the [14C]antipyrine volume and the interstitial volumes) remained unchanged. The wet weight-to-dry weight ratios of these lungs averaged 5.46 t 0.44. Left atria1 pressure studies (Fig. 4). Flows in these five studies averaged 63 ml/min. Moderate increases of left atria1 pressure (Pla) (6 * 0.5 Torr) had no signifi-
* ,
JO -
'
T
1T
.
T
PL
10 -
TIME (MIN\ 4. Left atria1 pressure studies. Increases in left atria1 pressure obtained during second period of pressure elevation resulted in statistically significant increases in Ppa, PVV, and filtration rate. control (FR, asterisk indicates significant difference from previous period.) EVV of 22Na+ and 1251-albumin increase in a manner similar to constant-flow and Ppa studies. FIG.
LUNGS
77
cant effect on Ppa or filtration. Greater increases in Pla (13 t 4 Torr) produced increases in Ppa, PVV, and the filtration rate. These parameters returned to normal values during the final control interval. Increases in the extravascular volumes of 1251-albumin and 22Na+ resembled those found in the isoflow and pulmonary artery studies. The wet weight-to-dry weight ratio averaged 5.83 t 0.97. DISCUSSION
The most striking characteristic of this lung preparation was the slow but progressive edema formation observed in all studies. As indicated above, the cause for edema formation in this and other isolated lungs remains uncertain. The constant-flow studies suggest that previous measurements of filtration, based on weight gain alone, are subject to error. We observed an initial decline in PVV that would diminish increases in lung weight related to passage of fluid into the tissues. Edema formation might, therefore, be underestimated by lung weight measurements. Furthermore, the increase in wet weight-to-dry weight ratios was very modest, when compared to controls, despite the presence of gross tracheal edema. This may also reflect the fact that the vasculature is diminished in drained, edematous lungs. It remains to be determined whether vascular contraction with edema formation occurs in isolated lungs of other species, or under other conditions. The decline of PVV occurred despite a fall in overall PVR. It would appear that early decreases in PVV are not accompanied by increases in PVR in this preparation. Although the PVV decline may reflect vascular compression by edema fluid, PVR failed to increase until gross edema formation was observed, and did not seem to be a sensitive parameter for the detection of transvascular filtration. The failure of PVV to decline further may be due, in part, to leakage of dextran blue into the tissues in the later stages of the study. Elevations of Pla and Ppa appear to produce a rapid expansion of the PVV followed by a more gradual increase in PVV that continued for as long as 15 min. Evidence has been obtained that the distension or recruitment of pulmonary vessels with increased hydrostatic pressure may occur over a relatively prolonged period of time (14). Although it has been proposed that pulmonary vascular filling is complete during the first 2 min (7, 8, 11), our study suggests that this may not be the case, and that the slow increase in lung weight used by others to detect filtration may include increases in the PVV. Modest elevations (6 Torr) in Pla produced small increases in mean Ppa, PVV, and filtration rate but none of these changes were significant. Greater elevations in Pla (13 Torr) produced significant increases in each of these parameters. Similarly, only large increases (10 Torr) in Ppa produced measurable increases in PVV and filtration rate. The decrease observed in the rate of filtration in the constant-flow studies may be due to a rise in interstitial pressure as edema accumulates. Alternatively, the per-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 11, 2019.
78
CHANG,
fused capillary surface area may diminish as edema becomes more severe. An apparent decline in filtration might also represent leakage of dextran blue into the interstitium in the later stages of edema formation. The extravascular volumes of distribution available to 22Na+ ggmTc-DTPA, and 1251-albumin increase progressively throughout these studies. The concentrations of each of these solutes in the edema fluid were close to those in the perfusion fluid. No change was found in the cellular volume, indicating that edema formation involved only the interstitial and/or alveolar compartments. Although the capillaries appeared to be permeable to solutes as large as albumin, a relatively constant portion of the extravascular space remained inaccessible to ggmTc-DTPA and (to an even greater extent) to 1251-albumin. It has been suggested that the interstitial matrix excludes larger molecules in a manner analo-
SILVERMAN,
AND
EFFROS
gous to Sephadex particles (10). The integrity of this volume appears to remain intact despite significant edema formation. Evidence has been provided that edema fluid may selectively enter the perivascular and peribronchiolar spaces (17), and that filling of the alveoli may occur after the bronchiolar wall has been breached (6). Bulk movement of protein and fluid may be largely confined to these channels, leaving the interstitial gel relatively unaltered. This research was supported by National Institutes of Health Grants HL-18606 and HL-00132 and by Grant GLAA562 from the A merican Heart Association, Greater Los Angeles Affiliate. We gratefully acknowledge the technical assistance of Ms. Kathy Wright and the fine secretarial work of Mrs. Abbey Rille. Received
28 February
1978; accepted
in final form 21 August
1978.
REFERENCES 1. Bij, G., A. HAUGE, AND G. NICOLAYSEN. Alveolar pressure and lung volume as determinants of net transvascular fluid filtration. J. Appl. Physiol.: Respirat. Environ. Exercise Physiol. 42: 476-482, 1977. 2. BORN, G. V. R. Acute edema in the isolated perfused lungs of rabbits. J. Physiol. London 124: 502-514, 1954. 3. BRODIE, B. B., E. Y. BEYER, J. ANEBROD, M. F. DUNNINE, Y. POROWSKA, AND J. M. STEELE. Use of n-acetyl 4-aminopyrine (NAAP) in measurement of total body water. Proc. Sot. Exp. Biol. Med. 77: 794-798, 1951. 4. EFFROS, R. M., AND F. P. CHINARD. The in vivo pH of the extravascular space of the lung. J. CZin. Invest. 48: 1983-1996, 1969. 5. GAAR, K. A., A. E. TAYLOR, L. J. OWENS, AND A. C. GUYTON. Pulmonary capillary pressure and filtration coefficient in the isolated perfused lung. Am. J. Physiol. 213: 910-914, 1967. 6. GEE, M. H., AND N. C. STAUB. Role of bulk fluid flow in protein permeability of the dog lung alveolar membrane. J. AppZ. Physiol.: Respirat. Environ. Exercise Physiol. 42: 144-149, 1977. 7. GOLDBERG, H. S., W. MITZNER, AND G. BATRA. Effect of transpulmonary and vascular pressures on rate of pulmonary edema formation. J. Appl. Physiol.: Respirat. Environ. Exercise PhysioZ. 43: 14-19, 1977. 8. HAUGE, A., G. Bij, AND B. A. WAALER. Interrelations between pulmonary liquid volumes and lung compliance. J. AppZ. PhysioZ. 38: 608-614, 1975. 9. KLOPPER, J. F., W. HAUSER, H. L. ATKINS, W. C. ECKELMAN, AND P. RICHARDS. Evaluation of ggmTc-DTPA for the measurement of glomerular filtration rate. J. NucZ. Med. 13: 107-110, 1972. 10. LAURENT, T. The structure and function of extracellular polysaccharides in connective tissue. In: CapiZZary Permeability, edited by C. Crone and N. A. Lassen. New York: Academic, 1970, p. 261-277. 11. LUNDE, P. K. M., AND B. A. WAALER. Transvascular fluid balance in the lung. J. Physiol. London 205: l-18, 1969. 12. NICOLAYSEN, G. Intravascular concentrations of calcium and magnesium ions and edema formation in isolated lungs. Acta
PhysioZ. Stand. 81: 325-339, 1971. 13. RYAN, J. W., U. SMITH, AND R. S. NIEMEYER. Angiotensin I: metabolism by plasma membrane of lung. Science 176: 64-66, 1972. 14. SARNOFF, S. J., AND E. BERGLUND. Pressure-volume characteristics and stress relaxation in the pulmonary vascular bed of the dog. Am. J. PhysioZ. 171: 238-244, 1952. 15. STACY, B. D., AND G. D. THOBURN. Chromium-51 ethylenediaminetetraacetate for estimation of glomerular filtration rate. Science 152: 1076-1077, 1966. 16. STAUB, N. C. Pulmonary edema. PhysioZ. Rev. 54: 678-811, 1974. 17. STAUB, N. C., H. NAGANO, AND M. L. PEARCE. Pulmonary edema in dogs, especially the sequence of fluid accumulation in the lungs. J. AppZ. PhysioZ. 22: 227-240, 1967. 18. TIERNEY, D. F., S. L. YOUNG, J. J. O’NEIL, AND M. ABE. Isolated perfused lung-substrate utilization. Federation Proc. 36: 161165, 1977. 19. TRAP-JENSEN, J., AND N. A. LASSEN. Capillary permeability for smaller hydrophilic tracers in exercising skeletal muscle in normal man and in patients with long-term diabetes mellitus. In: CapiZZary Permeability, edited by C. Crone and N. A. Lassen. New York: Academic, 1970, p. 135-152. 20. WAALER, B. A., AND P. AARSETH. Interstitial fluid and transcapillary fluid balance. In: Lung Liquids. Ciba Foundation Symposium, edited by C. J. Dickinson. London: Elsevier, 1975, p. 6571. 21. WEBB, H. H., AND D. F. TIERNEY. Experimental pulmonary edema due to intermittent positive pressure ventilation with high inflation pressures. Protection by positive end-expiratory pressure. Am. Rev. Respir. Dis. 110: 556-565, 1974. 22. WEISER, P. C., AND F. GRANDE. Estimation of fluid shifts and protein permeability during pulmonary edemagenesis. Am. J. PhysioZ. 226: 1028-1034, 1974. 23. WOOLVERTON, W., K. L. BRIGHAM, AND N. C. STAUB. Effect of continuous positive airway pressure breathing (CPAPB) on pulmonary fluid filtration and content in sheep (Abstract). PhysioZogist 16: 490, 1973.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 11, 2019.
