Experimental Evaluation of Reexpansion Pulmonary Edema Robert W. Sewell, M.D., John G. Fewel, B.A., Frederick L. Grover, M.D., a n d Kit V. Arom, M.D. ABSTRACT Reexpansion pulmonary edema following pneumothorax is clinically uncommon but occasionally life threatening. This study documents the functional and anatomical abnormalities that occur when a collapsed lung is reexpanded. Right pneumothorax was created through open tube thoracostomy in 30 goats. The animals were divided into six groups by duration of pneumothorax (24,48, or 72 hours) and technique of reexpansion (waterseal vs 10 cm H 2 0 suction). Arterial blood gases and alveolar-arterial oxygen tension difference (A-aDo,) were analyzed before pneumothorax and after reexpansion. Each lung was reexpanded for 2 hours, chest roentgenograms were obtained, and both lungs were removed. The left lung served as the control. Both lungs were checked for surfactant activity and pulmonary extravascular water volume (PEWV). Light and electron microscopy were also performed. Anatomical and functional changes were present in the reexpanded lung after relief of pneumothorax. Both increased time of collapse and suction reexpansion tended to correlate with increased PEWV, decreased surfactant and arterial PO,, and increased A-aDo,.

The treatment of pneumothorax by methods that rapidly reexpand the lung may result unexpectedly in life-threatening pulmonary edema. Carlson and co-workers [5] described this unusual complication in 1958. Since then, 12 additional cases have been reported, includFrom the Division of Cardiothoracic Surgery, The University of Texas Health Science Center at San Antonio and the Audie L. Murphy Veterans Administration Hospital, San Antonio, TX. We would like to express our gratitude to Dr. George A. Bannayan and Dr. David J. Jones for their technical assistance. Presented at the Twenty-fourth Annual Meeting of the Southern Thoracic Surgical Association, Nov 3-5, 1977, Marco Island, FL. Address reprint requests to Dr. Arom, Division of Cardiothoracic Surgery, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Dr, San Antonio, TX 78284.

ing 1 of our own [6, 11, 18, 20, 21, 25, 26, 281. There have been 2 deaths from acute respiratory insufficiency secondary to this form of unilateral pulmonary edema. Case Report A 28-year-old man was seen with a six-day history of dyspnea and right chest pain. Findings on physical examination were consistent with right pneumothorax, and a chest roentgenogram confirmed total collapse of the right lung (Fig 1A). An intercostal catheter was inserted and placed on 20 cm H,O suction. Immediately, the patient had a coughing paroxysm that produced moderate amounts of mucoid sputum. Over the next 2 hours hypotension and severe respiratory distress developed. A roentgenogram demonstrated massive unilateral pulmonary edema with partial recollapse of the lung (Fig 1B). The hypotension responded quickly to intravenous administration of colloid, but mechanical ventilation was necessary for the next 36 hours because of hypoxemia. Over the next ten days the pulmonary edema resolved and the lung expanded. The patient had been without sequelae for 18 months at the time of writing. There are several theories on the mechanism of this form of pulmonary edema. The factors thought to play a role are (1)the use of excessive suction for reexpansion, (2) decreased surfactant activity, and (3) extended duration of complete atelectasis. Since the severe form of pulmonary edema is encountered only rarely, there has been little stimulation to find an exact mechanism. Having experienced such a dramatic event in 1 of our patients, we resolved to examine the etiology experimentally. Materials and Methods Thirty adult female Spanish goats, averaging 26.1 kg in weight, underwent right pneumothorax through open tube thoracos-

126 0003-497517810026-0205$01.25 @ 1978 by Robert W. Sewell

127 Sewell et al: Reexpansion Pulmonary Edema

A Fig 1. (A)Admission roentgenogram shows total collapse of the right lung. ( B ) Massive unilateral pulmonary edema with partial collapse of the right lung is revealed in this chest roentgenogram made 2 hours after insertion of the chest tube.

B

weight. Ventilation was controlled with a Bird Mark VII ventilator connected to an endotube*The pressure in the sYstem was not allowed to exceed 20 cm H20. Each goat was placed in the supine position and a median sternotomy was performed, exposing tomy. Each goat was given acepromazine both lungs. The inferior pulmonary ligaments maleate, 3 mg per kilogram of body weight. were incised, and umbilical tapes were passed Then an open-end 18F Malecot catheter was around each hilum. The tapes were tied simulintroduced into the right pleural space, taneously to occlude all hilar structures comand a chest roentgenogram verified complete pletely. The time from initiation of anesthesia pneumothorax. Arterial blood samples were ob- to occlusion of both lungs did not exceed seven tained both before and after collapse by per- minutes. The animals were then killed by an cutaneous puncture of the carotid artery. The intraventricular injection of 10 ml of potassium animals were divided into six groups of 5 each chloride. Both lungs were transected proximal according to the duration of pneumothorax (24, to the umbilical tapes. Throughout the experi48, or 72 hours) and the method of reexpansion ment, the left (uncollapsed) lung served as the (10 cm H,O suction or underwater seal). After control. Quantitative measurements of pulmonary exthe collapse period, each lung was reexpanded by connecting the chest tube to a closed drain- travascular water volume ( P E W ) were made age unit, either with or without suction. Imme- using a modification of the postmortem diately prior to reexpansion, complete collapse gravimetric method of Pearce and associates was again documented by chest roentgenogram [8, 15, 231. A 100-gm sample of lung was and an arterial blood sample was obtained. homogenized along with 100 gm of distilled During the reexpansion period acepromazine water. Ten milliliters of the homogenate was weighed to obtain the specific gravity, and the maleate was used for sedation. A final roentgenogram and arterial blood remainder of the homogenate was centrifuged sample were obtained after 2 hours of reexpan- for 1 hour at 30,000 g and 4°C. Hemoglobin sion. The chest tube was then clamped and the concentrations were determined on eight sepaanimal anesthetized with Sodium Nembutal rate samples of the supernatant using the cya(pentobarbital), 30 mg per kilogram of body nomethemoglobin method. The entire homog-

128 The Annals of Thoracic Surgery Vol 26 No 2 August 1978

enate was recovered and dried at 85°C for 48 hours. The sample was allowed to cool and the dry weight measured. A 10-ml sample of peripheral venous blood was also measured for specific gravity, hemoglobin concentration, and dry weight. The following formulas [15] were used to calculate pulmonary blood weight (PBW) and PEWV.

pBw =

Hgb concentration of homogenate Hgb concentration of blood

spection revealed edematous fluid in the major bronchi and small bronchioles only in the group of lungs that had been collapsed for 72 hours and reexpanded with suction. Although all groups showed an increase in measured P E W (4.7 to 6.2 ml/lOO ml) compared with the control lung, there was no statistical difference between any of the groups ex-

Specific gravity of blood Specific gravity of homogenate

1.055*

P E W = Weight of lung - PBW - [Dry weight of lung - Dry weight of blood x PBW)] Wet weight of blood, L cept at 72 hours with suction reexpansion, when PEWV was 11.3 mUlOO ml (p < 0.01) (Fig 2). Decreased levels of pulmonary surfactant as measured by the bubble stability method were seen in the collapsed lung (p < 0.05 top < 0.01) except in the 24-hour and 72-hour waterseal groups. The mean decrease of the BSR ranged from 0.05 in the 24-hour waterseal group to 0.14 in the 48-hour suction group. The change in BSR in the 72-hour group, however, was less BSR = C squares of the diameters at 20 min than the 48-hour group (Fig 3 ) . 2 squares of the diameters at 0 min The most sensitive indicators of pulmonary edema in this study were arterial Po, and The arterial blood samples were analyzed A-aDO, measured 2 hours after reexpansion. for Po,, Pco,, and pH on an Instrumentation Laboratory blood gas analyzer system, and Major changes in these two factors were noted the aveolar-arterial oxygen tension difference even with small increases in P E W . The changes in arterial Po, (before pneumothorax vs (A-aDO,) was computed on each sample. after reexpansion) in the suction groups varied Small portions of each lung were fixed in 10% directly with duration of collapse. The maxiformaldehyde and stained with hematoxylin mum mean drop of 20.5 torr was at 72 hours. and eosin for light microscopy. Two percent glutaraldehyde was used to fix each specimen Decreases in Po, were significantly smaller in the waterseal groups, and no statistical for electron microscopy. difference was demonstrated between any of these nonsuction groups (Fig 4). Results As with Po, measurements, changes in All animals tolerated the collapsed lung very well, and at no time was there any roentgeno- A-aDo, were noted in all groups, the duration graphic evidence of major distortion of the of collapse again being the most important factor. The largest increase in A-aDo, was in mediastinum. Following reexpansion of the right lung, the 72-hour suction group (18.8 mm Hg), there were minimal roentgenographic changes with significantly smaller increases in the consistent with pulmonary edema. Gross in- waterseal groups (Fig 5). Light microscopical findings were consistent with the gross appearance of the lungs. The *Pulmonary hemoglobin correction factor. only group that showed intraalveolar edema Surfactant activity was tested using the bubble stability method described by Pattle and associates [14]. Polaroid micrographs of the suspended bubble preparations were obtained at zero and twenty minutes for each lung. The bubble stability ratio (BSR) was calculated by measuring the diameters of a minimum of 15 bubbles in each photomicrograph and using the formula:

129

Sewell et al: Reexpansion Pulmonary Edema

121

-.1" -

I

Key: Sdid=Suction Screen = Waterseal

L

6.2

E

t2.4

>

Key, Solid line=Suciion Dashed line-Waterseal

b

c

cc

I

$

5a

a

.-c

a

24 hrs 24 hrs

48 hrs

72 hrs

Fig 2. Changes in measured pulmonary extravascular water volume ( P E W ) of all groups. Note that the 72hour suction reexpansion group has a major increase in P E W of 11.3 mlilOO ml.

48hrs

72 hrs

Fig 4. Changes in arterial Po, before pneumothorax and after reexpansion in all groups. The maximum mean drop of 20.5 torr was noticed at 72 hours in the suction-reexpansion group.

Key: Solid line= Suction Dashed line = Waterseal Key: Solid =Suction Screen = Waterseal

0 15

niq

0 I4 to02

U

v)

m

a

010

005

0

24 hrs 24 hrs

48hrs

72 hrs

Fig 3. Changes in bubble stability ratio (BSR). The surfactant activity was decreased in all groups, but the decrease in the 72-hour group was significantly less than the 48-hour group. This is most likely due to the dilution of surface active material by the edematous fluid. There was no statistical difference between the size of initial bubbles in various experiments.

was the 72-hour suction group. In the lungs in this group there was no evidence of disruption of normal lung architecture; however, patchy microatelectasis was seen. Electron microscopy revealed some interesting changes not visible by light microscopy. The basement membranes were obviously thick-

48'hrs

7ihn

Fig 5. Changes in alveolar-arterial oxygen tension difference (n-aDo,) before pneumothorax and after reexdifference pansion in all groups. Note the high &a&, in the 48- and 72-hour suction reexpansion groups.

ened, and there was no interstitial edema despite extensive intraalveolar fluid (Fig 6 ) . Again, these changes were found only in the 72-hour suction reexpansion group. Comment This experiment deals with the problem of pneumothorax and the pulmonary edema that may result from reexpansion of the lung. While very few cases of this complication have been reported, it is probable that mild degrees of

130 The Annals of Thoracic Surgery Vol 26

No 2 August 1978

A

B

Fig 6 . ( A )Electron micro,graph of the control lunx slzowing normal basement membrane (BM) and interstitium (IN) and no edematous fluid in t h e alveoli (AV). ( B ) Electron micro,cyap/i ._. of a lung from the 72-hour suction reexpansion group showing t'jzlckening of t h e basement membrane, a/veo/ar edema, and no edema in the interstitial tissue. (Original magnification X2,400.)

In 1875 Foucart [lo] proposed a mechanism for reexpansionpulmonary edema. H~ believed that during the period Of collapseg the lung was subjected to anoxic injury sufficient to damage the pulmonary capillaries, thus increasing their permeability. After this insult, fluid could rapidly move into the interstitial and alveolar spaces when the lung was reexpanded. The autoregulatory mechanisms of the pulmonary circulation provide for total, or near total, cessation of pulmonary blood flow during periods of atelectasis [21. It is this loss of pulmonary flow that is thought to provide the anoxic insult [7]. This theory was supported by Humphreys and Berne [Ill, Carlson [5], Ratliff [181, and Odelowo [13] and their associates with respect to reexpansion pulmonary edema after pneumothorax. They pointed out that both Barach and colleagues [l] and Warren and associates [27] had shown that even a short period of anoxia increases capillary permeability. Humphreys and Berne [ l l ] also were aware that rapid reexpansion of a collapsed lung results in an increase in pulmonary capillary perfusion pressure and pulmonary blood flow. These factors aid in the movement of fluid from the capillaries into the interstitial and alveolar spaces

I

edema occur frequently. It also is reasonable to assume that the etiology is the same as that of reexpansion pulmonary edema following the rapid removal of a large pleural effusion [26]. This phenomenon was first described by Pinault [16] in 1853 in a patient who had 3 liters of fluid removed rapidly by thoracentesis. In 1901 Riesman [19] published a review outlining the clinical signs. He noted that the problem seldom occurred unless at least 2 liters of fluid was rapidly removed. He also recognized that a symptom-free period of several minutes to a few hours usually preceded the onset of copious sputum production and respiratory distress. In most instances the symptoms cleared rapidly; however, on occasion they lasted as long as 24 to 48 hours. Finally, he noted that the cause of death in patients who died of this problem was acute obstructive respiratory insufficiency secondary to massive unilateral pulmonary edema.

WI.

131 Sewell et al: Reexpansion Pulmonary Edema

Sutnick and Soloff [24] demonstrated that surfactant activity in the atelectatic lung was markedly lower than that of the normal lung. This finding is confirmed by our data. They observed a gradual increase in surface tension with duration of collapse. Trapnell and Thurston [25] and Sautter and associates [211 suggested that reexpansion pulmonary edema was caused by loss of surfactant activity. With the resultant high surface tension, the lung tends to recollapse, requiring greater than normal negative intrapleural pressures to maintain expansion. This tendency to recollapse has been documented in patients following rapid removal of a chronic pneumothorax [3, 5, 18, 191. We likewise observed recollapse in our patient as noted in Figure 1. Pattle and co-workers [14] obtained these findings experimentally and demonstrated that low levels of pulmonary surfactant resulted in movement of fluid into reexpanded alveoli. In our experimental model we noted important decreases in surfactant activity in all collapsed lungs. The decrease in surfactant activity in the 72-hour group, however, was less than expected and could be related to the dilution of surface active material by the pulmonary extravascular lung water. Dilution is greater in the 72-hour group than in the other groups. Ziskind [28] and Childress [6] and their associates believed that the application of excessive intrathoracic suction was a major cause of pulmonary edema. However, suction was used in only 6 of 17 reported cases of reexpansion pulmonary edema following pneumothorax. Miller [12] demonstrated that suction was necessary to produce pulmonary edema in primates with total pneumothorax of three days’ duration. However, extremely high negative intrapleural pressures (100 mm Hg or 135 cm H,O suction) were used for reexpansion. Carlson and co-workers [5] were able to demonstrate pulmonary edema in rabbits following reexpansion of a pneumothorax with syringe aspiration. Our results document these findings, in that significant increases in PEWV were seen in both the waterseal and the suction groups. Also, the increase in PEWV in each group is statistically correlated (by regression analysis) with the increase in A-aDOP differences. It must be pointed out, though, that

the greatest change in PEWV was in the 72-hour suction reexpansion group. We have shown in this experiment that unilateral pulmonary edema following reexpansion of a chronic pneumothorax is associated with several factors. The duration of collapse of the lung is of primary importance. When collapse was present for less than 72 hours, the measured increase in pulmonary extravascular water volume remained small. This may explain why the complication is not encountered more frequently. In his series of 376 spontaneous pneumothoraces in 307 patients, Brooks [41 reported that 70% of the patients sought medical attention within the first 24 hours and only 10% consulted a physician one week or more after the apparent onset of symptoms. Thus, most patients are treated before important changes occur in the microphysiology of the lung. We also demonstrated that changes occurred in the alveolar-capillary basement membrane after 72 hours of collapse, rendering it more permeable. This injury probably is a direct result of anoxia since pulmonary circulation as well as ventilation is interrupted in atelectatic areas. The

abrupt reestablishment of circulation into a damaged capillary bed results in dislocation of fluid into the alveolar spaces. Pulmonary edema may result from catheter drainage, either with or without suction, as in our experiment, or with needle aspiration alone. The important factor is the removal of large volumes of air or fluid in a short period. The natural formation of high negative intrapleural pressure in the acutely expanded lung provides an additional mechanism by which protein-rich fluid leaves the vascular space. Negative intrapleural pressures occur due to low levels of surfactant, with resultant high alveolar surface tension, causing a tendency to recollapse [14, 171. The combination of the greater than normal capillary permeability and the increased negative intrathoracic pressure results in rapid transudation of fluid out of the capillaries. The use of additional suction in reexpanding the lung with the resultant increase in negative intrapleural pressure presumably caused a rise in the mean capillary flow and perfusion pressure within that lung. This may have contributed to an increase in filtration pressure and

132 The Annals of Thoracic Surgery Vol 26

No 2 August 1978

fluid transudation in our experiment but is not always present in clinically significant reexpansion pulmonary edema. Treatment must be based on the degree of respiratory embarrassment. Stabilization of the pulmonary capillary membrane may be aided by the use of intravenous steroids. The administration of colloid solutions and diuretics may also retard edema accumulation [19]. In severe cases the use of continuous positive airway pressure may offset the effects of high alveolar surface tension by helping to reverse the transcapillary movement of fluid 191. Prevention is the most effective means of dealing with this problem. Delays in reexpansion beyond three days should be avoided when possible. However, patients seen with a chronically collapsed lung can be treated safely by aspiration or catheter drainage if large volumes are not removed rapidly. Frequent aspirations of small amounts of air or fluid over a 24or 48-hour period should provide satisfactory reexpansion without pulmonary edema.

References 1. Barach AL, Martin J, Eckman M: Positive pressure

2.

3.

4. 5.

6.

7.

8.

respiration and its application to the treatment of acute pulmonary edema. Ann Intem Med 12:754, 1938 Bergofsky EH: Mechanisms underlying vasomotor regulation of regional pulmonary blood flow in normal and disease states. Am J Med 57:378, 1974 Bernstein A, Waqaruddin M, Shah M: Management of spontaneous pneumothorax using a Heimlich flutter valve. Thorax 28:386, 1973 Brooks JW: Open thoracotomy in the management of spontaneous pneumothorax. Ann Surg 177:798, 1973 Carlson RI, Classen KL, Gollan F, et al: Pulmonary edema following the rapid reexpansion of a totally collapsed lung due to pneumothorax: a clinical and experimental study. Surg Forum 9:367,1958 Childress ME, Moy G, Mottram M: Unilateral pulmonary edema resulting from treatment of spontaneous pneumothorax. Am Rev Respir Dis 104:119, 1971 Dei Poli G, Andolfi F, Teggia I'M: Spontaneous idiopathic pneumothorax in 115 young subjects. Panminerva Med 16207, 1974 Erdmann AJ 111, Vaughan TR Jr, Brigham KL, et al: Effect of increased vascular pressure on the lung fluid balance in unanesthetized sheep. Circ Res 37271, 1975

9. Fishman AP: Pulmonary edema; the waterexchanging function of the lung. Circulation 46:390, 1972 10. Foucart (1875): Cited by Riesman [19] 11. Humphreys RL, Beme AS: Rapid reexpansion of pneumothorax, a cause of unilateral pulmonary edema. Radiology 96:509, 1970 12. Miller WC: Experimental pulmonary edema following the reexpansion of pneumothorax. Am Rev Respir Dis 108:664, 1973 13. Odelowo EO, Calhoun T, Kurtz LH, et al: Spontaneous pneumothorax. J Natl Med Assoc 66:111, 1974 14. Pattle RE, Claireaux AE, Davies PA, et al: Inability to form a lung-lining film as a cause of the respiratory-distress syndrome in the newborn. Lancet 2:469, 1962 15. Pearce ML, Yamashita J, Beazell J: Measurement of pulmonary edema. Circ Res 16:482, 1965 16. Pinault (1853): Cited by Riesman [19] 17. Radford EP: Mechanical stability of the lung. Arch Environ Health 6:134, 1963 18. Ratliff JL, Chavez CM, Jamchuk A, et al: Reexpansion pulmonary edema. Chest 64:654, 1973 19. Riesman D: Albuminous expectoration following thoracocentesis. Am J Med Sci 123:620, 1902 20. Saini GS: Unilateral pulmonary edema after drainage of spontaneous pneumothorax. Br Med J 1:615, 1974 21. Sautter RD, Dreher WH, MacIndoe JH, et al: Fatal pulmonary edema and pneumonitis after reexpansion of chronic pneumothorax. Chest 60:399, 1971 22. Schneeberger EE, Karnovsky MJ: The influence of intravascular fluid volume o n the permeability of newborn and adult mouse lungs to ultrastructural protein tracers. J Cell Biol49:319, 1971 23. Selinger SL, Bland RD, Demling RH, et al: Distribution volumes of ""1 albumin, I4C sucrose, a n d " T 1in sheep lung. J Appl Physiol39:773,1975 24. Sutnick AI, Soloff LA: Surface tension reducing activity and atelectatic human lung. Am J Med 35:31, 1963 25. Trapnell DH, Thurston JGB: Unilateral pulmonary oedema after pleural aspiration. Lancet 1:1367, 1970 26. Waqaruddin M, Bernstein A: Reexpansion pulmonary oedema. Thorax 30:54, 1975 27. Warren MF, Peterson DK, Drinker CK: The effects of heightened negative pressure in the chest, together with further experiments upon anoxia in increasing the flow of lung lymph. Am J Physiol 137641, 1942 28. Ziskind MM, Weill H, George RA: Acute pulmonary edema following the treatment of spontaneous pneumothorax with excessive negative intrapleural pressure. Am Rev Respir Dis 92:632, 1965