Unilateral pulmonary edema in rabbits after reexpansion of collapsed lung JANET PAVLIN AND FREDERICK W. CHENEY, JR. Department of Anesthesiology, University of Washington School of Medicine, Seattle, Washington 98195
PAVLIN,JANET, AND FREDERICK W. CHENEY, JR. Unilateral pulmonary edema in rabbits after reexpansion of collapsed lung. J. Appl. Physiol. : Respirat. Environ. Exercise Physiol. 46(l): 31-35, 1979. -The effects of the mode of reinflation and of the duration of prior collapse on the development of unilateral pulmonary edema following reexpansion of collapsed lung were studied in a rabbit model simulating the human syndrome of “reexpansion pulmonary edema.” The right lungs of rabbits were maintained in an atelectatic state for 0.5 h to 8 days, by injection of air into the pleural space. Reexpansion was achieved in 2 h by application of positive pressure to the airway while a chest tube was connected to underwater seal, or by application of negative pressure (-20 to - 100 Torr) to a screened window in the parietal pleura. The lung surface pressures we actually applied by the two methods are not known. Animals were then killed and pulmonary edema was determined by wet-to-dry weight ratios. The incidence of unilateral pulmonary edema increased as the duration of prior collapse was increased (85% after 7-8 days; 17% after 3 days; and 0% after 0.5 h) when reinflated with - 100 Torr applied to the pleural window. Although the incidence was less, it also occurred following the use of pleural window pressures less negative than -100 Torr, and after reinflation by positive airway pressure. pneumothorax;
atelectasis
PULMONARY EDEMA is a syndrome observed in humans, characterized by the development of unilateral pulmonary edema in a lung that has been rapidly reinflated following a variable period of collapse resulting from pleural effusion or pneumothorax. Typically this response occurs in patients when a lung collapsed for 3 days or more is rapidly reexpanded by negative pressure applied to the pleural space (4, 5, 12, 17, 18). The ensuing unilateral pulmonary edema is associated with variable degrees of hypoxia and hypotension, sometimes requiring intubation and mechanical ventilation for a period of several days. Within 5-7 days, edema has usually resolved, although two deaths have been reported in association with this sequence of events (13, 17). Although the number of reported cases is small, the actual incidence of this syndrome is unknown as most cases are undoubtedly not documented in the literature. This study was designed to determine whether 1) the duration of prior collapse or 2) positive or negative pressure used to reinflate the lung influenced the incidence or severity of edema. The informa-
REEXPANSION
0161-7567/79/0000-0000$01.25
Copyright
0 1979 the American
tion to be gained might then allow one to predict the likelihood of this syndrome developing in a clinical situation, and possibly ascertain methods of lung reinflation that would be least likely to precipitate this problem. MATERIALS
AND
METHODS
New Zealand White rabbits, 3-4 kg, of both sexes were used as the experimental animal. Collapse of one lung was induced by injection of 80 ml of air into the right pleural cavity under fluoroscopic control. Injections were then repeated daily to maintain a state of near total collapse for 0.5 h, 3 days, or 7-8 days. Prior to reinflation, animals were anesthetized with intraperitoneal pentobarbital, 35-45 mg/kg, with supplements given as often as necessary to prevent movement. Animals were intubated and breathed pure oxygen spontaneously from a nonrebreathing circuit via an endotracheal tube. Initial attempts to reinflate lungs by use of a pediatric chest tube resulted in frequent obstruction of the tubing within the thorax. To obviate this problem a circular window, about 3 cm in diameter, was created in the right lateral chest wall by resecting two or three ribs. A fiberglass screen was then sewn over the defect to prevent extrusion of the underlying lung during reinflation‘ A large plastic chest tube, 2 cm in diameter with a wide flange, was glued over the defect and an airtight seal obtained. The lung wasthen reinflated in one of two ways: 1) negative pressure of - 20 -40, or -100 Torr applied continuously to the chest tube while the animals breathed pure oxygen spontaneously or 2) intermittent positive-pressure ventilation with oxygen. In the latter instance a Harvard pump was set to deliver a tidal volume of 15 ml/kg, at a rate of approximately 18/min, with manual hyperinflations up to an airway pressure of 25 cm H,O applied every 20 min. In these animals, the chest tube was connected to an underwater seal. In all instances, reinflation was allowed to proceed for 2 h, at which time animals were killed by the injection of a 60-mg intravenous bolus of pentobarbital. While the heart was still beating, the animals were exsanguinated by transection of the inferior vena cava, and the lungs were then excised. Each lung was divided into three sections: two sections were used for duplicate measurements of lung edema by determination of wet-
Physiological
Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on October 11, 2018. Copyright © 1979 American Physiological Society. All rights reserved.
31
32
J. PAVLIN
to-dry weight ratios of lung; the third section was used to determine blood content of lung. Lung sections were weighed immediately after excision, then dried in a vacuum desiccator at 80°C for several days until the weight was stable. The water content of lung was determined by subtraction of the dry weight from the wet weight. The weights were corrected for the weight of blood contained within the third section of lung as determined by the hemoglobin technique (15). Wet-todry weight ratios were expressed as H,O content of blood-free lung/dry weight of blood-free lung as recommended by Staub (15). In the first group of 40 animals, the effect of the mode of reinflation was assessed. After 7-8 days of collapse, lungs were 1) not reinflated (n = 7); 2) reinflated with negative pressures applied to the pleural window (Pwin) of -20 (n = 8), -40 (n = S), or -100 (n = 13) Torr; or3) reinflated with positive-pressure ventilation with a chest tube connected to underwater seal (n = 8). In a second group of 14 animals, the effects of the duration of prior collapse was examined. Lungs reinflated with Pwin of -100 Torr negative pressure after 0.5 h or 3 days were compared to those reinflated after 7-8 days of collapse. In a third group of four animals, an attempt was made to simulate bronchial obstruction in one lobe of the collapsed lung. After 0.5 h of collapse, a no. 5 SwanGanz catheter introduced through a tracheal incision was threaded into one of the branches of the right main bronchus and the balloon was inflated so as to occlude the bronchus distally. Reinflation was then accomplished with -100 Torr negative pressure as in other experiments. In a fourth group of eight animals (normal. controls), with no prior surgery or pneumothorax, rabbits were killed in the usual manner and wet-to-dry weight ratios of normal control lungs determined. In two animals, pulmonary artery, right atrial, and right ventricular pressures were measured using a no. 5 Swan-Ganz catheter inserted via an external jugular vein into the right heart chambers and pulmonary artery. RESULTS
The water content of blood-free lung/dry weight of blood-free lung of eight normal control right lungs was 3 78 + - 0.31 (SD). In evaluating individual lungs in all of the other experimental groups, a ratio > 2 SD above the mean of the control group (B4.39) was considered abnormally high (P c 0.05) and indicative of the presence of pulmonary edema. These lungs were obviously edematous on gross examination and oozed frothy edema fluid from the cut surface. The effects of the mode of reinflation are illustrated in Fig. 1 and Table 1. The incidence of unilateral edema on the side of prior collapse was greatest if lungs were reinflated when Pwin was -100 Torr, but occurred in a number of animals when lesser amounts of negative pressure, or positive-pressure ventilation were used. The mean values of lung water content are significantly different from the normal control lungs when reinflated with Pwin of -40 or - 100 Torr negative pressure, or
AND
F. W. CHENEY,
JR.
positive pressure at the airway; there is no significant difference between the water con tent of lungs reinflated by any of these three methods. Because of the lesser severity of pulmonary edema and the small number of animals that developed it in the group reinflated with Pwin of -20 Torr, the mean values of this latter group differ significantly from the -lOO-Torr group, but not from the control lungs. Two animals reinflated with Pwin of -40 Torr developed measurable edema of the left lung as well as the right. In both cases, the right lungs were still considerably more edematous than the left (water content of right lung of 5.04 vs. 4.61 in the left lung in one animal, and 7.08 vs. 5.34 in the other animal). Edema of the left lung may have been due to aspiration of edema fluid from the right lung into the left lung. Alternatively, the left lung may have also been partially compressed by the right-sided pneumothorax present before reinflation; edema might then have bccurred by the same mechanism causing edema of the right lung. It is also possible that mild cardiac failure might have occurred as the result of hypoxia and hypercarbia resulting in small increases in lung water bilaterally. However, the predominance of edema fluid in the right lung suggests 1001
’
2 80 c 5
R= Right
I
lung
L= Left
n=l3
lung
-
1 n=8
-20
I
torr
torr
n=8
R
L
R
Control 0
L
R
Collapse with no reinflation
L
R
L
-40
R - 100 torr
L
R
L
Positive airway pressure
Reinflation pressures 1. Effect of mode of reinflation. Extravascular lung water content/dry weight of blood-free lung (g/g dry lung) > 4.39 = edematous lung. Right lungs collapsed 7-8 days before reinflation. FIG.
100~
I
R= Right L= Left
lung lung
Con tfol
n=13
l/2
hour
3 days
7-8 days
Duration of prior collapse 2. Effect of duration of prior collapse. Extravascular lung water content/dry weight of blood-free lung (g/g dry lung) > 4.39 = edematous lung. Right lungs reinflated with negative chest tube pressures of -100 Torr. FIG.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on October 11, 2018. Copyright © 1979 American Physiological Society. All rights reserved.
REEXPANSION
PULMONARY
33
EDEMA
that reexpansion was still the major cause of edema of the right lung. The effects of the duration of prior collapse are shown in Fig. 2 and Table 1. After 8 days of collapse, unilateral edema developed in 85% of animals. After 3 days, however, edema still occurred in 17% of animals, although the mean value for water content of this group is not significantly different from the control group. In a third group of animals with bronchial obstruction present during reinflation after 0.5 h of collapse, there was no significant difference between the water content of the obstructed lobes and the other nonobstructed lobes of the same side, the opposite lung or the normal control lungs. These results are shown in Table 2. 1. Effect of reinflation
TABLE
Duration of Right Lung Collapse
No. of Lungs
Conditions
on lung water content
Reinflation
Pressure
Right
A. Control
8
None
None
B. Atelectasis
7
7-8 days
None
C. Atelectasis
13
7-8 days
- 100 Torr
D. Atelectasis
8
7-8 days
-40 Torr
E. Atelectasis
8
7-8 days
-20 Torr
F. Atelectasis
6
3 days
-100 Torr
G. Atelectasis
8
0.5 h
- 100 Torr
H. Atelectasis
8
7-8 days
+25 cmH,O airwayg
* Values are means t SD. t P< lungs (Student’s t test for unpaired data). left lungs (Student’s paired t test). ventilation.
-
Extravascular Lung Water Content/Dry Wt Blood-Free Lung,* g/g dry lung Left
3.82 20.28 3.72 +0.18 5.03 +0.84$$ 4.71 +1.25t$ 4.08 t0.66$ 4.00 t0.66 3.80 kO.19 4.59 tl.05tS
3.79 20.16 4.00 kO.14 3.76 to.39 4.23 kO.57 3.70 to.21 3.66 kO.32 3.90 kO.15 4.07 1~0.23
0.05 compared to control $ P < 0.05 compared to Q With positive-pressure
TABLE 2. Effect of bronchial obstruction on water content of right lungs reinflated after 0.5-h collapse Extravascular Condition
of Right Lung
Lung Water Content/Dry Free Lung,* g/g dry lung
No. of Lungs Right
A. Controlt B. Reinflated, unobstructed C. Reinflated, obstructed
8 3
3.82 + 0.29 3.60 + 0.15
4
3.69 + 0.44
Wt
Blood-
Left
3.79 ~fr 0.16 3.76 + 0.14$
* Values are means f SD. t Normal control values from previous experiments. $ The left lungs, tabulated with group B, were the normal contralateral lungs with no prior history of collapse or reinflation. RV
PA
Suction + - 1ootorr
In two animals pulmonary artery, right ventricular, and right atria1 pressures were measured before and after onset of -100 Torr Pwin. There was an initial transient fall in pulmonary artery pressure lasting only a few seconds with the onset of negative pressure in the chest tube, followed by a return to control levels (Fig.
3) . DISCUSSION
Our study confirms
the experimental
work of Miller in monkeys. In contrast to Miller who found that edema occurred in all lungs reinflated after 3 days of collapse, we detected edema only occasionally after 3 days of collapse, but after 7-8 days of collapse, pulmonary edema was present in almost all animals. In the present study, the amount of edema has been quantitated by measurement of excess water content of the edematous lungs in contrast to previous works in which gross anatomic and histological criteria were used to assess the presence of edema (2, 8). From our data, it would appear that the frequency and presumably the severity of “reexpansion pulmonary edema” are related to the duration of prior collapse, with the frequency of edema increasing as the duration of collapse is prolonged up to 8 days. We found that edema occurred most frequently when high levels of negative pressure were applied to a screened pleural window to reinflate the lung. It also occurred, though with a lesser frequency, when lower amounts of negative pressure or positive airway pressure was used to achieve reinflation. It is possible that the more frequent occurrence of edema after “high negative-pressure reinflation” is related to a more rapid and complete reexpansion of lung. However, we did not measure the rate of extent of changes in lung volume or the pressures actually applied to the surface of the lung. The mechanism of the development of unilateral pulmonary edema remains unclear. Normally we assume that a balance of forces exists at the level of the pulmonary microvasculature whereby fluid is retained within the vasculature according to the Starling hypothesis. This balance of forces is presumably weighted slightly in favor of a net transudation of fluid out of the vasculature, which is then removed by the lymphatic system and ultimately returned to the vascular compartment. Accumulation of water in the lung may occur due to an increase in the net rate of fluid flux out of the vascular compartment and/or a reduction in its rate of removal from the interstitial compartment. Considering the former possibility, fluid flux into lung interstitium may be increased by either an increase in the hydro-
(8) who was a bl e t o produce this syndrome
Flush + FIG. 3. Hemodynamic effect of negative-pressure reinflation. RV, right ventricle; PA, pulmonary artery.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on October 11, 2018. Copyright © 1979 American Physiological Society. All rights reserved.
J. PAVLIN
34
static pressure gradient across the pulmonary microvasculature, or a reduction in the effective colloid osmotic pressure gradient as might occur with altered vascular permeability to protein macromolecules. As pulmonary artery pressures were not altered appreciably by lung reexpansion, it is unlikely that intravascular pressures were altered at the sites of fluid exchange. However, a reduction in perivascular pressure surrounding the microvasculature could provide a mechanism whereby transvascular hydrostatic pressure gradients would be altered to produce increased transudation of fluid into lung interstitium. Such a reduction in perivascular pressure might occur if extra-alveolar vessels were lying in a potential space within the lung. As noncompliant lungs were forcibly reexpanded to conform to the configuration of the chest cavity, perivascular spaces might be expanded, causing a reduction in perivascular pressure and resultant transudation of fluid along a hydrostatic gradient. Levine and Johnson (6) have demonstrated increased recoil pressures when chronically collapsed lungs are reinflated in vitro, suggesting that pleural pressures might become very negative when such lungs are reinflated in vivo. In our experiments, negative pleural and interstitial pressures generated by increased lung recoil pressure must have been limited by the amount of negative pressure that the chest wall could support. Experiments by Cavagna et al. (3) suggest that the rabbit has a relatively compliant chest wall an.d that static pressures achieved in the pleural space are limited to about -10 Torr with the chest wall relaxed. However, the fact that different results were obtained with Pwin of -20 and -100 Torr suggests that the reexpansion process was not completely determined by any limits which may have been imposed by chest wall elasticity. Although we do not know what pleural pressures were attained during lung reexpansion, one may speculate that initially, when air was rapidly evacuated from the right chest, pleural pressure must have been at least -9 Torr in accordance with Cavagna’s data (3) and probably was considerably more negative due to active contraction of the respiratory muscles in our spontaneously breathing animals. In our experiments, the cardiac impulse was initially very markedly shifted to the right with the onset of negative-pressure reinflation, lending support to the supposition that right-sided pleural pressures were initially significantly reduced. Levine and Johnson (6) found that complete reexpansion of previously collapsed lungs required static pressures approximately 18 Torr greater than those required to inflate normal lungs. One might speculate, therefore, that in our experiments, pleural pressures were at least -9 to - 18 Torr, depending on the balance of forces between lung recoil pressure, respiratory muscle activity, and chest wall compliance. Additionally, liquid pleural and interstitial pressures, which are the relevant determinants of fluid flux in the lung, may be slightly more negative than surface pressures; the latter are probably more important in determining lung volume (9). Furthermore, pleural and interstitial pressures
AND
F. W. CHENEY,
JR.
might vary considerably throughout the chest, perhaps attaining very negative values in areas of persistent (7). Stalcup and Mellins (14) have demonatelectasis strated that relatively sma 11 reductions of pleural pressure in dogs using a body plethysmograph resulted in significant increases in lung water in the presence of an intact chest wall, suggesting that one need not invoke very large changes in pleural pressure to explain fluid accumulation in the lung. One might predict, from their data, that reducing mean pleural pressure from 0 to -10 Torr would result in approximately a 30% increase in lung water over a 2-h period at constant pulmonary vascular and colloid osmotic pressure in dogs. An alternative explanation of the mechanism of reexpansion edema is that vascular permeability to macromolecules and water may be increased either as the result of vascular “injury” sustained during prolonged collapse, and/or injury or “stretching” of vascular pores during lung reexpansion. The failure to detect any increase in lung water in lungs that were collapsed for 8 days but not reexpanded suggests that there was no change in vascular permeability prior to reexpansion, permeability is unless one postulates that increased masked in the presence of atelectasis by reduction in perfusion of the collapsed lung (10) and/or decreased vascular surface area available for fluid exchange. It seems unlikely that such changes would be of a magnitude that they exactly counterbalance the effects of altered permeability. Reexpansion, however, might alter permeability by “stretching” preexisting endothelial pores, or causing the formation of new Yents” in vascular endothelium. If the interstitial fluid pressure is truly reduced when chronically collapsed lung is reinflated, interference with lymphatic drainage might be an additional causative factor. Staub (15) has proposed, on the basis of measurement of pressure in terminal lymphatics, that fluid moves into lymphatics along a hydrostatic gradient of pressure. Reduction in interstitial pressure might reduce this gradient and thereby slow the rate of fluid removal via lymphatics. It is conceivable that decreased surfactant might play a role in the formation of pulmonary edema by altering the surface tensi .on in such a way as to augment tendency of atelectatic alveoli to stay “collapsed,” thereby increasing lung recoil pressure. Levine and Johnson (6), however, found that lungs collapsed for 8 days continued to exhibit hysteresis when air pressurevolume curves were performed consistent with preservation of surfactant activity. Their work casts some doubt on the relative importance of changes in surface forces in the lung as a cause of “reexpansion pulmonary edema.” Our study demonstrates the potentially harmful effects of rapidly reinflating a lung which has been totally collapsed for a period of several days by the presence of pneumothorax. It is possible that si .milar forces may be operant during reexpansion of focal areas of ate1 .ectasis, although this circumstance has not been specifically examined. BQ) and Vaage (1) have demonstrated a 1214% increase of extravascular lung water in cat lungs
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on October 11, 2018. Copyright © 1979 American Physiological Society. All rights reserved.
REEXPANSION
PULMONARY
35
EDEMA
allowed to become partially atelectatic by continuous ventilation with small tidal volumes. The increases in lung water were partially reversible by hyperinflating lungs in vivo to reinflate areas of atelectasis. This suggests a possible role of regional atelectasis in the generation of focal areas of edema. In summary, it is apparent that some change occurs within the lung during prolonged atelectasis, which strongly predisposes to the development of pulmonary
edema when the lung is reinflated. Whether this is caused by a change in pulmonary vascular permeability, and/or the mechanical properties of lung, remains to be explained. The authors gratefully acknowledge the technical assistance of Rodney Gronka and Ed Chappelle. This study was supported by Anesthesia Research Training Grant GMO-1160-12 and Pulmonary SCOR Center Grant HL-14152. Received
20 May 1978; accepted in final form 2 August
1978.
REFERENCES 1. Bs, G., AND J. VAAGE. Lung water volume in atelectasis. BibZ. Anat. 12: 376-380, 1973. 2. CARLSON, R. J., K. L. CLASSEN, F. GOLLAN, W. G. GOBBEL, D. E. SHERMAN, AND R. 0. CRISTENSEN. Pulmonary edema following the rapid re-expansion of a totally collapsed lung due to a pneumothorax. Surg. Forum 9: 367-371, 1959. 3. CAVAGNA, G. A., E. J. STEMMLER, AND A. B. DUBOIS. Resistance to atelectasis. J. AppZ. PhysioZ. 22: 441-452, 1967. 4. CHILDRESS, M. E., G. MOY, AND M. MOTTRAM. Unilateral pulmonary edema resulting from treatment of spontaneous pneumothorax. Am. Rev. Respir. Dis. 104: 119-121, 1971. 5. HUMPHREYS, R. I,., AND A. S. BERNE. Rapid re-expansion of pneumothorax. RadioLogy 96: 509-512, 1970. 6. LEVINE, B., AND R. P. JOHNSON. Effects of atelectasis on pulmonary surfactant and quasi-static lung mechanics. J. AppZ. Physiol. 20: 859-864, 1970. 7. MACKLEM, P. M., B. MURPHY, AND B. ENG. The forces applied to the lung in health and disease. Am. J. Med. 57: 371-377, 1974. 8. MILLER, W.C. Experimental pulmonary edema following reexpansion of pneumothorax. Am. Rev. Respir. Dis. 108: 664-666, 1973. 9. PERMUTT, S., AND P. CALDINI. Tissue pressure and fluid dynamics of the lungs. Federation Proc. 35: 1876-1880. 1976.
10. PETERS, R. M., AND A. Roos. The effects of atelectasis upon pulmonary blood flow in the lung. J. Thorac. Surg. 24: 389-396, 1952. 11. RATCLIFFE, J. L., C. M. CHAVEZ, A. JAMECHUK, J. E. FORESTNER, AND J. H. CONN. Re-expansion pulmonary edema. Chest 64: 654656, 1973. 12. SAUTTER, R. D., W. H. DREHER, J. H. MACINDOE, W. 0. MYERS, AND G. E. MAGNIN. Fatal pulmonary edema and pneumonitis after re-expansion of chronic pneumothorax. Chest 60: 399-401, 1971. 13. SHANAHAN, M. X., I. MONK, AND H. J. RICHARDS. Unilateral pulmonary oedema following re-expansion of pneumothorax. Anaesth. Intensive Care 3: 19-30, 1975. 14. STALCUP, S. A., AND R. B. MELLINS. Mechanical forces producing pulmonary edema in acute asthma. N. EngZ. J. Med. 297: 592596, 1977. 15. STAUB, N. C. Pulmonary edema. Physiol. Rev. 54: 678-811, 1974. 16. STECKEL, R. J. Unilateral pulmonary edema after pneumothorax. N. EngZ. J. Med. 289: 621-622, 1973. 17. TRAPNELL, D. H., AND J. G. B. THURSTON. Unilateral pulmonary oedema after pleural aspiration. Lancet 1: 1367-1369, 1970. 18. WAQARUDDIN, M., AND A. BERNSTEIN. Re-expansion pulmonary oedema. Thorax 30: 54-60. 1975.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on October 11, 2018. Copyright © 1979 American Physiological Society. All rights reserved.
PAVLIN,JANET, AND FREDERICK W. CHENEY, JR. Unilateral pulmonary edema in rabbits after reexpansion of collapsed lung. J. Appl. Physiol. : Respirat. Environ. Exercise Physiol. 46(l): 31-35, 1979. -The effects of the mode of reinflation and of the duration of prior collapse on the development of unilateral pulmonary edema following reexpansion of collapsed lung were studied in a rabbit model simulating the human syndrome of “reexpansion pulmonary edema.” The right lungs of rabbits were maintained in an atelectatic state for 0.5 h to 8 days, by injection of air into the pleural space. Reexpansion was achieved in 2 h by application of positive pressure to the airway while a chest tube was connected to underwater seal, or by application of negative pressure (-20 to - 100 Torr) to a screened window in the parietal pleura. The lung surface pressures we actually applied by the two methods are not known. Animals were then killed and pulmonary edema was determined by wet-to-dry weight ratios. The incidence of unilateral pulmonary edema increased as the duration of prior collapse was increased (85% after 7-8 days; 17% after 3 days; and 0% after 0.5 h) when reinflated with - 100 Torr applied to the pleural window. Although the incidence was less, it also occurred following the use of pleural window pressures less negative than -100 Torr, and after reinflation by positive airway pressure. pneumothorax;
atelectasis
PULMONARY EDEMA is a syndrome observed in humans, characterized by the development of unilateral pulmonary edema in a lung that has been rapidly reinflated following a variable period of collapse resulting from pleural effusion or pneumothorax. Typically this response occurs in patients when a lung collapsed for 3 days or more is rapidly reexpanded by negative pressure applied to the pleural space (4, 5, 12, 17, 18). The ensuing unilateral pulmonary edema is associated with variable degrees of hypoxia and hypotension, sometimes requiring intubation and mechanical ventilation for a period of several days. Within 5-7 days, edema has usually resolved, although two deaths have been reported in association with this sequence of events (13, 17). Although the number of reported cases is small, the actual incidence of this syndrome is unknown as most cases are undoubtedly not documented in the literature. This study was designed to determine whether 1) the duration of prior collapse or 2) positive or negative pressure used to reinflate the lung influenced the incidence or severity of edema. The informa-
REEXPANSION
0161-7567/79/0000-0000$01.25
Copyright
0 1979 the American
tion to be gained might then allow one to predict the likelihood of this syndrome developing in a clinical situation, and possibly ascertain methods of lung reinflation that would be least likely to precipitate this problem. MATERIALS
AND
METHODS
New Zealand White rabbits, 3-4 kg, of both sexes were used as the experimental animal. Collapse of one lung was induced by injection of 80 ml of air into the right pleural cavity under fluoroscopic control. Injections were then repeated daily to maintain a state of near total collapse for 0.5 h, 3 days, or 7-8 days. Prior to reinflation, animals were anesthetized with intraperitoneal pentobarbital, 35-45 mg/kg, with supplements given as often as necessary to prevent movement. Animals were intubated and breathed pure oxygen spontaneously from a nonrebreathing circuit via an endotracheal tube. Initial attempts to reinflate lungs by use of a pediatric chest tube resulted in frequent obstruction of the tubing within the thorax. To obviate this problem a circular window, about 3 cm in diameter, was created in the right lateral chest wall by resecting two or three ribs. A fiberglass screen was then sewn over the defect to prevent extrusion of the underlying lung during reinflation‘ A large plastic chest tube, 2 cm in diameter with a wide flange, was glued over the defect and an airtight seal obtained. The lung wasthen reinflated in one of two ways: 1) negative pressure of - 20 -40, or -100 Torr applied continuously to the chest tube while the animals breathed pure oxygen spontaneously or 2) intermittent positive-pressure ventilation with oxygen. In the latter instance a Harvard pump was set to deliver a tidal volume of 15 ml/kg, at a rate of approximately 18/min, with manual hyperinflations up to an airway pressure of 25 cm H,O applied every 20 min. In these animals, the chest tube was connected to an underwater seal. In all instances, reinflation was allowed to proceed for 2 h, at which time animals were killed by the injection of a 60-mg intravenous bolus of pentobarbital. While the heart was still beating, the animals were exsanguinated by transection of the inferior vena cava, and the lungs were then excised. Each lung was divided into three sections: two sections were used for duplicate measurements of lung edema by determination of wet-
Physiological
Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on October 11, 2018. Copyright © 1979 American Physiological Society. All rights reserved.
31
32
J. PAVLIN
to-dry weight ratios of lung; the third section was used to determine blood content of lung. Lung sections were weighed immediately after excision, then dried in a vacuum desiccator at 80°C for several days until the weight was stable. The water content of lung was determined by subtraction of the dry weight from the wet weight. The weights were corrected for the weight of blood contained within the third section of lung as determined by the hemoglobin technique (15). Wet-todry weight ratios were expressed as H,O content of blood-free lung/dry weight of blood-free lung as recommended by Staub (15). In the first group of 40 animals, the effect of the mode of reinflation was assessed. After 7-8 days of collapse, lungs were 1) not reinflated (n = 7); 2) reinflated with negative pressures applied to the pleural window (Pwin) of -20 (n = 8), -40 (n = S), or -100 (n = 13) Torr; or3) reinflated with positive-pressure ventilation with a chest tube connected to underwater seal (n = 8). In a second group of 14 animals, the effects of the duration of prior collapse was examined. Lungs reinflated with Pwin of -100 Torr negative pressure after 0.5 h or 3 days were compared to those reinflated after 7-8 days of collapse. In a third group of four animals, an attempt was made to simulate bronchial obstruction in one lobe of the collapsed lung. After 0.5 h of collapse, a no. 5 SwanGanz catheter introduced through a tracheal incision was threaded into one of the branches of the right main bronchus and the balloon was inflated so as to occlude the bronchus distally. Reinflation was then accomplished with -100 Torr negative pressure as in other experiments. In a fourth group of eight animals (normal. controls), with no prior surgery or pneumothorax, rabbits were killed in the usual manner and wet-to-dry weight ratios of normal control lungs determined. In two animals, pulmonary artery, right atrial, and right ventricular pressures were measured using a no. 5 Swan-Ganz catheter inserted via an external jugular vein into the right heart chambers and pulmonary artery. RESULTS
The water content of blood-free lung/dry weight of blood-free lung of eight normal control right lungs was 3 78 + - 0.31 (SD). In evaluating individual lungs in all of the other experimental groups, a ratio > 2 SD above the mean of the control group (B4.39) was considered abnormally high (P c 0.05) and indicative of the presence of pulmonary edema. These lungs were obviously edematous on gross examination and oozed frothy edema fluid from the cut surface. The effects of the mode of reinflation are illustrated in Fig. 1 and Table 1. The incidence of unilateral edema on the side of prior collapse was greatest if lungs were reinflated when Pwin was -100 Torr, but occurred in a number of animals when lesser amounts of negative pressure, or positive-pressure ventilation were used. The mean values of lung water content are significantly different from the normal control lungs when reinflated with Pwin of -40 or - 100 Torr negative pressure, or
AND
F. W. CHENEY,
JR.
positive pressure at the airway; there is no significant difference between the water con tent of lungs reinflated by any of these three methods. Because of the lesser severity of pulmonary edema and the small number of animals that developed it in the group reinflated with Pwin of -20 Torr, the mean values of this latter group differ significantly from the -lOO-Torr group, but not from the control lungs. Two animals reinflated with Pwin of -40 Torr developed measurable edema of the left lung as well as the right. In both cases, the right lungs were still considerably more edematous than the left (water content of right lung of 5.04 vs. 4.61 in the left lung in one animal, and 7.08 vs. 5.34 in the other animal). Edema of the left lung may have been due to aspiration of edema fluid from the right lung into the left lung. Alternatively, the left lung may have also been partially compressed by the right-sided pneumothorax present before reinflation; edema might then have bccurred by the same mechanism causing edema of the right lung. It is also possible that mild cardiac failure might have occurred as the result of hypoxia and hypercarbia resulting in small increases in lung water bilaterally. However, the predominance of edema fluid in the right lung suggests 1001
’
2 80 c 5
R= Right
I
lung
L= Left
n=l3
lung
-
1 n=8
-20
I
torr
torr
n=8
R
L
R
Control 0
L
R
Collapse with no reinflation
L
R
L
-40
R - 100 torr
L
R
L
Positive airway pressure
Reinflation pressures 1. Effect of mode of reinflation. Extravascular lung water content/dry weight of blood-free lung (g/g dry lung) > 4.39 = edematous lung. Right lungs collapsed 7-8 days before reinflation. FIG.
100~
I
R= Right L= Left
lung lung
Con tfol
n=13
l/2
hour
3 days
7-8 days
Duration of prior collapse 2. Effect of duration of prior collapse. Extravascular lung water content/dry weight of blood-free lung (g/g dry lung) > 4.39 = edematous lung. Right lungs reinflated with negative chest tube pressures of -100 Torr. FIG.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on October 11, 2018. Copyright © 1979 American Physiological Society. All rights reserved.
REEXPANSION
PULMONARY
33
EDEMA
that reexpansion was still the major cause of edema of the right lung. The effects of the duration of prior collapse are shown in Fig. 2 and Table 1. After 8 days of collapse, unilateral edema developed in 85% of animals. After 3 days, however, edema still occurred in 17% of animals, although the mean value for water content of this group is not significantly different from the control group. In a third group of animals with bronchial obstruction present during reinflation after 0.5 h of collapse, there was no significant difference between the water content of the obstructed lobes and the other nonobstructed lobes of the same side, the opposite lung or the normal control lungs. These results are shown in Table 2. 1. Effect of reinflation
TABLE
Duration of Right Lung Collapse
No. of Lungs
Conditions
on lung water content
Reinflation
Pressure
Right
A. Control
8
None
None
B. Atelectasis
7
7-8 days
None
C. Atelectasis
13
7-8 days
- 100 Torr
D. Atelectasis
8
7-8 days
-40 Torr
E. Atelectasis
8
7-8 days
-20 Torr
F. Atelectasis
6
3 days
-100 Torr
G. Atelectasis
8
0.5 h
- 100 Torr
H. Atelectasis
8
7-8 days
+25 cmH,O airwayg
* Values are means t SD. t P< lungs (Student’s t test for unpaired data). left lungs (Student’s paired t test). ventilation.
-
Extravascular Lung Water Content/Dry Wt Blood-Free Lung,* g/g dry lung Left
3.82 20.28 3.72 +0.18 5.03 +0.84$$ 4.71 +1.25t$ 4.08 t0.66$ 4.00 t0.66 3.80 kO.19 4.59 tl.05tS
3.79 20.16 4.00 kO.14 3.76 to.39 4.23 kO.57 3.70 to.21 3.66 kO.32 3.90 kO.15 4.07 1~0.23
0.05 compared to control $ P < 0.05 compared to Q With positive-pressure
TABLE 2. Effect of bronchial obstruction on water content of right lungs reinflated after 0.5-h collapse Extravascular Condition
of Right Lung
Lung Water Content/Dry Free Lung,* g/g dry lung
No. of Lungs Right
A. Controlt B. Reinflated, unobstructed C. Reinflated, obstructed
8 3
3.82 + 0.29 3.60 + 0.15
4
3.69 + 0.44
Wt
Blood-
Left
3.79 ~fr 0.16 3.76 + 0.14$
* Values are means f SD. t Normal control values from previous experiments. $ The left lungs, tabulated with group B, were the normal contralateral lungs with no prior history of collapse or reinflation. RV
PA
Suction + - 1ootorr
In two animals pulmonary artery, right ventricular, and right atria1 pressures were measured before and after onset of -100 Torr Pwin. There was an initial transient fall in pulmonary artery pressure lasting only a few seconds with the onset of negative pressure in the chest tube, followed by a return to control levels (Fig.
3) . DISCUSSION
Our study confirms
the experimental
work of Miller in monkeys. In contrast to Miller who found that edema occurred in all lungs reinflated after 3 days of collapse, we detected edema only occasionally after 3 days of collapse, but after 7-8 days of collapse, pulmonary edema was present in almost all animals. In the present study, the amount of edema has been quantitated by measurement of excess water content of the edematous lungs in contrast to previous works in which gross anatomic and histological criteria were used to assess the presence of edema (2, 8). From our data, it would appear that the frequency and presumably the severity of “reexpansion pulmonary edema” are related to the duration of prior collapse, with the frequency of edema increasing as the duration of collapse is prolonged up to 8 days. We found that edema occurred most frequently when high levels of negative pressure were applied to a screened pleural window to reinflate the lung. It also occurred, though with a lesser frequency, when lower amounts of negative pressure or positive airway pressure was used to achieve reinflation. It is possible that the more frequent occurrence of edema after “high negative-pressure reinflation” is related to a more rapid and complete reexpansion of lung. However, we did not measure the rate of extent of changes in lung volume or the pressures actually applied to the surface of the lung. The mechanism of the development of unilateral pulmonary edema remains unclear. Normally we assume that a balance of forces exists at the level of the pulmonary microvasculature whereby fluid is retained within the vasculature according to the Starling hypothesis. This balance of forces is presumably weighted slightly in favor of a net transudation of fluid out of the vasculature, which is then removed by the lymphatic system and ultimately returned to the vascular compartment. Accumulation of water in the lung may occur due to an increase in the net rate of fluid flux out of the vascular compartment and/or a reduction in its rate of removal from the interstitial compartment. Considering the former possibility, fluid flux into lung interstitium may be increased by either an increase in the hydro-
(8) who was a bl e t o produce this syndrome
Flush + FIG. 3. Hemodynamic effect of negative-pressure reinflation. RV, right ventricle; PA, pulmonary artery.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on October 11, 2018. Copyright © 1979 American Physiological Society. All rights reserved.
J. PAVLIN
34
static pressure gradient across the pulmonary microvasculature, or a reduction in the effective colloid osmotic pressure gradient as might occur with altered vascular permeability to protein macromolecules. As pulmonary artery pressures were not altered appreciably by lung reexpansion, it is unlikely that intravascular pressures were altered at the sites of fluid exchange. However, a reduction in perivascular pressure surrounding the microvasculature could provide a mechanism whereby transvascular hydrostatic pressure gradients would be altered to produce increased transudation of fluid into lung interstitium. Such a reduction in perivascular pressure might occur if extra-alveolar vessels were lying in a potential space within the lung. As noncompliant lungs were forcibly reexpanded to conform to the configuration of the chest cavity, perivascular spaces might be expanded, causing a reduction in perivascular pressure and resultant transudation of fluid along a hydrostatic gradient. Levine and Johnson (6) have demonstrated increased recoil pressures when chronically collapsed lungs are reinflated in vitro, suggesting that pleural pressures might become very negative when such lungs are reinflated in vivo. In our experiments, negative pleural and interstitial pressures generated by increased lung recoil pressure must have been limited by the amount of negative pressure that the chest wall could support. Experiments by Cavagna et al. (3) suggest that the rabbit has a relatively compliant chest wall an.d that static pressures achieved in the pleural space are limited to about -10 Torr with the chest wall relaxed. However, the fact that different results were obtained with Pwin of -20 and -100 Torr suggests that the reexpansion process was not completely determined by any limits which may have been imposed by chest wall elasticity. Although we do not know what pleural pressures were attained during lung reexpansion, one may speculate that initially, when air was rapidly evacuated from the right chest, pleural pressure must have been at least -9 Torr in accordance with Cavagna’s data (3) and probably was considerably more negative due to active contraction of the respiratory muscles in our spontaneously breathing animals. In our experiments, the cardiac impulse was initially very markedly shifted to the right with the onset of negative-pressure reinflation, lending support to the supposition that right-sided pleural pressures were initially significantly reduced. Levine and Johnson (6) found that complete reexpansion of previously collapsed lungs required static pressures approximately 18 Torr greater than those required to inflate normal lungs. One might speculate, therefore, that in our experiments, pleural pressures were at least -9 to - 18 Torr, depending on the balance of forces between lung recoil pressure, respiratory muscle activity, and chest wall compliance. Additionally, liquid pleural and interstitial pressures, which are the relevant determinants of fluid flux in the lung, may be slightly more negative than surface pressures; the latter are probably more important in determining lung volume (9). Furthermore, pleural and interstitial pressures
AND
F. W. CHENEY,
JR.
might vary considerably throughout the chest, perhaps attaining very negative values in areas of persistent (7). Stalcup and Mellins (14) have demonatelectasis strated that relatively sma 11 reductions of pleural pressure in dogs using a body plethysmograph resulted in significant increases in lung water in the presence of an intact chest wall, suggesting that one need not invoke very large changes in pleural pressure to explain fluid accumulation in the lung. One might predict, from their data, that reducing mean pleural pressure from 0 to -10 Torr would result in approximately a 30% increase in lung water over a 2-h period at constant pulmonary vascular and colloid osmotic pressure in dogs. An alternative explanation of the mechanism of reexpansion edema is that vascular permeability to macromolecules and water may be increased either as the result of vascular “injury” sustained during prolonged collapse, and/or injury or “stretching” of vascular pores during lung reexpansion. The failure to detect any increase in lung water in lungs that were collapsed for 8 days but not reexpanded suggests that there was no change in vascular permeability prior to reexpansion, permeability is unless one postulates that increased masked in the presence of atelectasis by reduction in perfusion of the collapsed lung (10) and/or decreased vascular surface area available for fluid exchange. It seems unlikely that such changes would be of a magnitude that they exactly counterbalance the effects of altered permeability. Reexpansion, however, might alter permeability by “stretching” preexisting endothelial pores, or causing the formation of new Yents” in vascular endothelium. If the interstitial fluid pressure is truly reduced when chronically collapsed lung is reinflated, interference with lymphatic drainage might be an additional causative factor. Staub (15) has proposed, on the basis of measurement of pressure in terminal lymphatics, that fluid moves into lymphatics along a hydrostatic gradient of pressure. Reduction in interstitial pressure might reduce this gradient and thereby slow the rate of fluid removal via lymphatics. It is conceivable that decreased surfactant might play a role in the formation of pulmonary edema by altering the surface tensi .on in such a way as to augment tendency of atelectatic alveoli to stay “collapsed,” thereby increasing lung recoil pressure. Levine and Johnson (6), however, found that lungs collapsed for 8 days continued to exhibit hysteresis when air pressurevolume curves were performed consistent with preservation of surfactant activity. Their work casts some doubt on the relative importance of changes in surface forces in the lung as a cause of “reexpansion pulmonary edema.” Our study demonstrates the potentially harmful effects of rapidly reinflating a lung which has been totally collapsed for a period of several days by the presence of pneumothorax. It is possible that si .milar forces may be operant during reexpansion of focal areas of ate1 .ectasis, although this circumstance has not been specifically examined. BQ) and Vaage (1) have demonstrated a 1214% increase of extravascular lung water in cat lungs
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on October 11, 2018. Copyright © 1979 American Physiological Society. All rights reserved.
REEXPANSION
PULMONARY
35
EDEMA
allowed to become partially atelectatic by continuous ventilation with small tidal volumes. The increases in lung water were partially reversible by hyperinflating lungs in vivo to reinflate areas of atelectasis. This suggests a possible role of regional atelectasis in the generation of focal areas of edema. In summary, it is apparent that some change occurs within the lung during prolonged atelectasis, which strongly predisposes to the development of pulmonary
edema when the lung is reinflated. Whether this is caused by a change in pulmonary vascular permeability, and/or the mechanical properties of lung, remains to be explained. The authors gratefully acknowledge the technical assistance of Rodney Gronka and Ed Chappelle. This study was supported by Anesthesia Research Training Grant GMO-1160-12 and Pulmonary SCOR Center Grant HL-14152. Received
20 May 1978; accepted in final form 2 August
1978.
REFERENCES 1. Bs, G., AND J. VAAGE. Lung water volume in atelectasis. BibZ. Anat. 12: 376-380, 1973. 2. CARLSON, R. J., K. L. CLASSEN, F. GOLLAN, W. G. GOBBEL, D. E. SHERMAN, AND R. 0. CRISTENSEN. Pulmonary edema following the rapid re-expansion of a totally collapsed lung due to a pneumothorax. Surg. Forum 9: 367-371, 1959. 3. CAVAGNA, G. A., E. J. STEMMLER, AND A. B. DUBOIS. Resistance to atelectasis. J. AppZ. PhysioZ. 22: 441-452, 1967. 4. CHILDRESS, M. E., G. MOY, AND M. MOTTRAM. Unilateral pulmonary edema resulting from treatment of spontaneous pneumothorax. Am. Rev. Respir. Dis. 104: 119-121, 1971. 5. HUMPHREYS, R. I,., AND A. S. BERNE. Rapid re-expansion of pneumothorax. RadioLogy 96: 509-512, 1970. 6. LEVINE, B., AND R. P. JOHNSON. Effects of atelectasis on pulmonary surfactant and quasi-static lung mechanics. J. AppZ. Physiol. 20: 859-864, 1970. 7. MACKLEM, P. M., B. MURPHY, AND B. ENG. The forces applied to the lung in health and disease. Am. J. Med. 57: 371-377, 1974. 8. MILLER, W.C. Experimental pulmonary edema following reexpansion of pneumothorax. Am. Rev. Respir. Dis. 108: 664-666, 1973. 9. PERMUTT, S., AND P. CALDINI. Tissue pressure and fluid dynamics of the lungs. Federation Proc. 35: 1876-1880. 1976.
10. PETERS, R. M., AND A. Roos. The effects of atelectasis upon pulmonary blood flow in the lung. J. Thorac. Surg. 24: 389-396, 1952. 11. RATCLIFFE, J. L., C. M. CHAVEZ, A. JAMECHUK, J. E. FORESTNER, AND J. H. CONN. Re-expansion pulmonary edema. Chest 64: 654656, 1973. 12. SAUTTER, R. D., W. H. DREHER, J. H. MACINDOE, W. 0. MYERS, AND G. E. MAGNIN. Fatal pulmonary edema and pneumonitis after re-expansion of chronic pneumothorax. Chest 60: 399-401, 1971. 13. SHANAHAN, M. X., I. MONK, AND H. J. RICHARDS. Unilateral pulmonary oedema following re-expansion of pneumothorax. Anaesth. Intensive Care 3: 19-30, 1975. 14. STALCUP, S. A., AND R. B. MELLINS. Mechanical forces producing pulmonary edema in acute asthma. N. EngZ. J. Med. 297: 592596, 1977. 15. STAUB, N. C. Pulmonary edema. Physiol. Rev. 54: 678-811, 1974. 16. STECKEL, R. J. Unilateral pulmonary edema after pneumothorax. N. EngZ. J. Med. 289: 621-622, 1973. 17. TRAPNELL, D. H., AND J. G. B. THURSTON. Unilateral pulmonary oedema after pleural aspiration. Lancet 1: 1367-1369, 1970. 18. WAQARUDDIN, M., AND A. BERNSTEIN. Re-expansion pulmonary oedema. Thorax 30: 54-60. 1975.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on October 11, 2018. Copyright © 1979 American Physiological Society. All rights reserved.
