- Email: [email protected]
The effects of ventilation with positive end-expiratory pressure on the bronchial circulation
269
Respiration Physiology (1986) 66, 269-278 Elsevier
THE EFFECTS OF VENTILATION WITH POSITIVE END-EXPIRATORY PRESSURE ON THE BRONCHIAL CIRCULATION*
SHARON S. CASSIDY and MICHAEL S. HAYNES Laboratoryfor Cardiopulmonary Research and the Pulmonary Research Division of the Department of lnternal Medicine, University of Texas Southwestern Medical School Dallas, 75233, U.S.A.
Abstract. We studied the effects of ventilation with 10 cm H20 PEEP for 2 h in dogs with temporary unilateral pulmonary arterial occlusion (TUPAO) on bronchial blood flow to the occluded lung using the microsphere dispersion technique. We found that blood flow to the occluded left lung in dogs was 9.9 ml/min ( 0 . 1 2 2 m l ' m i n - t ' g - I ) . Within 30min following the addition of 10cm H20 PEEP blood flow fell by 70-80 ~ (to 2.3 ml/min) caused both by a 3-fold decrease in vascular conductance and a 25 % fall in systemic blood pressure. The reduction in left bronchial blood flow persisted for at least 2 h. We conclude from these data that ventilation with PEEP in the presence of pulmonary artery occlusion has a severe, persistent adverse effect on bronchial blood flow. This reduction in bronchial blood flow is beyond what can be explained by the changes in airway pressure. The additional increase in bronchial vascular resistance may be caused by the increase in lung volume, by reflex bronchial vasoconstriction, or by release of mediators locally. Dog Lung injury
Microsphere PEEP
TUPAO
Aside from its critical role in various types of congenital heart disease, bronchial circulation has been implicated as being potentially important in the formation of pulmonary edema caused by increased permeability (Pietra and Magno, 1978) and in the sequelae of pulmonary artery occlusion (Ellis et al., 1951; William and Towbin, 1955; Parker and Smith, 1957). Johnson et al. (1981) recently studied the effects of a temporary unilateral pulmonary arterial occlusion (TUPAO) on a group of dogs and found that by histological assessment as well as by measurement of lung tissue volume (C2H 2 method), the lung distal to the TUPAO developed edema and injury. The extent of the injury was enhanced by ventilation with 10 cm H20 PEEP. We reasoned that Accepted for publication 22 August 1986 * A preliminary report was presented to the Federation of American Societies of Experimental Biology, April, 1979 in Dallas, Texas (Haynes etal., 1979) 0034-5687/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)
270
S.C. CASSIDY AND M.S. HAYNES
an adverse effect of PEEP on bronchial blood flow in the setting of TUPAO could have contributed to lung injury, and the purpose of the present study was to determine the effects of PEEP on bronchial arterial blood flow during TUPAO using the microsphere dispersion technique. The microsphere method for measuring organ blood flow has not been applied often to the bronchial circulation because of the potential leakage of microspheres from the systemic circulation into venous blood with subsequent entrapment in the lungs via the pulmonary circulation. Because pulmonary arterial blood flow to one lung was occluded in this study, the potential for contamination with microspheres which leaked through the systemic circulation was eliminated for the occluded lung.
Methods
Twelve mongrel dogs of either sex weighing 18-22 kg were studied under pentobarbital anesthesia, 25 to 30 ml/kg, i.v. They were ventilated with a large animal sinusoidal pump ventilator at 15 ml/kg via an endotracheal tube at a rate sufficient to maintain medal Pco2 at approximately 40 mm Hg. During left thoracotomy a catheter was placed in the left atrial appendage for subsequent injection of radionuclide microspheres. The chest was closed tightly and residual air was aspirated from the pleural space. Two catheters were placed in the femoral artery, one for sampling during microsphere injections, the other for monitoring medal pressure. In addition, catheters were advanced to the right atrium and pulmonary artery via femoral veins. A balloon-occlusion catheter was placed in the left pulmonary artery and was inflated to occlude blood flow. Thirty minutes following occlusion, measurements of blood flow distributions were made without PEEP (IPPV). In nine dogs additional measurements were made 30, 60, and 120 min following the addition of 10 cm H20 PEEP. In 3 dogs measurements were made at similar intervals but without adding PEEP. Arterial blood gases, pH, and pressure recordings were obtained prior to each blood flow measurement. The technique for measuring organ blood flow using dispersion of radionuclide microspheres was adapted from the original description of Rudolph and Heymann (1967) using the reference sampling adaption ofArchie et al. (1973) and Domenech et al. (1969). Radionuclides used in random order were 14ICe, 125I, SSSr, and 46Sc. A 3-ml mixture containing 500 000 25-/~m latex spheres suspended in Tween was injected via the left atrial catheter and flushed with 30 ml warmed saline over 30 sec. The time to clear the dead space of the blood withdrawal system which pumped at 10 ml/min was 20 sec and was independent of medal pressure. After a 20 sec delay, blood was also sampled from the pulmonary artery for 180 sec to detect microsphere spillover from the systemic vascular bed. At the termination of the experiment, the animals were killed by exsanguination, organs were removed, and connective tissue was dissected away. The lungs were divided into right and left, upper and lower lobes, and these tissues were minced and placed in preweighed vials. The vials containing tissue and blood were
BRONCHIAL BLOOD FLOW DURING PEEP
271
reweighed and placed in a scintillation counter for determining the gamma emission at preselected channels corresponding to the peak KEV emission of each nuclide. Spillover of emission from one nuclide into channels detecting the peak KEV of the other nuclides was subtracted according to the methods previously described by Rudolph and Heymann (1967).
Statistical analyses.
Data are expressed as means + SEM. Data were compared with a one way analysis of variance for repeated measures (Zar, 1974). When the F value indicated a significant non-random difference (P < 0.05) among repeated measures, Newman-Keuls multiple comparison procedure was used to determine significance of difference between pairs of means (Zar, 1974). P < 0.05 was chosen as the level reflecting significance.
Results
Bronchial arterial blood flow during unilateral pulmonary artery occlusion and PEEP (fig. 1). Following left TUPAO bronchial blood flow measured 9.9 ml/min. Blood flow was higher to the left lower lobe compared to the left upper lobe, 0.135 vs 0.110 ml. rain - t .g - ,, and this trend continued following PEEP. Blood flow to the left lung (with the left pulmonary artery occluded) fell by 70% to 72% within 30 min after placing the animal on 10 cm H20 PEEP. Through the next 1.5 h bronchial blood flow continued to fall until it was 84-85% below pre-PEEP levels. Blood flow to the occluded left lung in three dogs that were not placed on PEEP did not change significantly (falling by 20%) over a period of 1.5 h following the fu'st measurement of blood flow 30 rain post occlusion.
Bronchial arterial driving pressure during unilateral pulmonary arterial occlusion and PEEP (fig. I). The driving pressure for bronchial blood flow should be the difference between systemic m e d a l pressure and either pulmonary venous or alveolar pressure depending o • o • I
Ls.
RIght to',,,m"k~:~e Right t#ppe¢Iot~ Left I o ~ r lobe Left ~oper I o ~ tS.E.
o
S~em~c orfery
3OO
(SA)
• .~A-wq~dge
E 2
100 qcz
0 ~e 30 PEEP
60
90
t20
I Pre
1 30
I 60
I 90
I 120
0
PEEP TIME DURING tOcm H20 PEEP (rain)
Fig. 1. Bronchialarterial blood flows and bronchial arterial driving pressures followingleft pulmonary arterial occlusionbefore and during2 h ventilationwith 10 cm H20 PEEP.
272
s.c. CASSIDY AND M.S. HAYNES
TABLE 1 Pressures that may influence bronchial blood flowduring temporary unilateral pulmonaryartery occlusion before and durillg ventilation with 10 ¢m H20 PEEP. Pulmonary arterial pressure (cm H20)
Pulmonary wedge pressure (cm H20)
Systemic arterial pressure (mm Hg)
Airway pressure (cm H20)
Before PEEP
28.3 _+2.5
9.6 4. 1.7
148 4- 13
8.3 _+0.9
During PEEP 30 min
28.6 _4-5.5
13.3" + 2.5
88* 4. 8
16.4" _+1.4
60 rain
30.6 4_4.6
13.3 4_2.9
102 +_11
18.1 +0.8
120 min
28.9 +3.7
11.9 +3.1
95 4. 7
18.6 4.1.0
Values are means + SEM, n = 6. * P < 0.05 pre-PEEP vs 30 rain post-PEEP. All comparisons between 30 min and 120 rain were not significant.
on which system (pulmonary venous or alveolar) has the highest critical closing pressure to bronchial outflow. Absolute pulmonary arterial, pulmonary wedge, systemic arterial and airway pressures are listed in table 1. Airway pressures rose from 8 cm H 2 0 to 16.4-18.6 cm H 2 0 with PEEP and pulmonary arterial mean pressure did not change significantly. Depending on which measure was used, driving pressure fell either by 49% or 46% initially with 10 cm H 2 0 PEEP. With further passage of time driving pressure did not continue to fall but was restored slightly ( 3 - 6 % ) by virtue of the restoration of systemic arterial pressure.
Left bronchial vascular resistance during TUPA O and PEEP (fig. 2).
Bronchial vascular resistance to the occluded left lung increased 2-3 fold with the addition of 10 cm H 2 0 PEEP. With passage of time left bronchial vascular resistance rose further so that after 2 h the resistance was 4 times greater than pre-PEEP values. The conductance of the left bronchial vascular channels fell to 1/2 to 1/3 of pre-PEEP levels after 2 h of ventilation with PEEP.
Distribution of microspheres to mixed venous blood. See Appendix. Arterial and venous pH and blood gases (table 2).
A mild metabolic acidosis developed over the 2 h period during T U P A O and PEEP. Pao2 remained above 100 mm Hg while mixed venous Po, fell and Pco2 increased both probably reflecting a low output state.
BRONCHIAL BLOOD FLOW DURING PEEP
273
g
'c E "F i
Ns
1.0
¢
-,..i. . . . . .I. . . . .
E E
i
Ns
5"
os 33
o ~Z
3
i
== I
I
I
!
I
l
I
I
0
30
6O
120
0
3O
6O
f20
0
I
Time After 10 cm H20 PEEP (rain)
Fig. 2. Left bronchial arterial resistance and conductance followingleft TUPAO before and during 2 h ventilation with 10 cm H20 PEEP. TABLE 2 Arterial and mixed venous blood gas tensions and pH
pH
Before PEEP
Pco2
Po2
Arterial
Mixed venous
Arterial
Mixed venous
Arterial
Mixed venous
7.41 ± 0.03
7.39 _+0.02
33.7 _+2.5
38.3 ± 3.5
100 _+ 5
39 ±3
7.29* _+0.04
30.3 _+6.7
43.0* _+8.2
113 ± 13
32* ±7
Following 10 cm H20 PEEP 30 rain 7.35* _+0.05 60 rain
7.33 ± 0.02
7.28 ± 0.02
32.0 ± 6.9
41.7 _+7.6
113 _+12
36 4-4
120 min
7.301. _+0.02
7.231" _+0.01
34.0 _+5.2
51.3' _+8.1
117 _+23
291" _+4
Means± SE; * P < 0.05 pre-PEEP vs 30 min post-PEEP; 1.P < 0.05 post-PEEP 30 min vs 120 min. L u n g weights. Averaged post m o r t e m wrights of the lung were: left upper lobe, 31.8 +_ 1.9g; left lower lobe, 47.5 _+ 2 . 7 g ; left lung 81.0 _+ 4 . 7 g ; fight upper lobe, 42.7 _+ 1.89 g; fight lower lobe, 52.1 +_ 3 . 0 g ; a n d right lung, 94.8 +_ 4.2g.
Discussion Bronchial blood flow to the entire left lung averaged 9.9 ml/min. This corresponds reasonably with m e a s u r e m e n t s made with flow meters in lungs receiving their n o r m a l
274
s.c. CASSIDY AND M.S. HAYNES
complement of pulmonary arterial flow. Bruner and Schmidt (1947) reported right posterior bronchial blood flow (60~ of all right bronchial flow) was 4.8 ml/min, Horisberger and Rodbard (1960) reported bilateral bronchial blood flows of 12.1 ml/min, and Baile et al. (1982) reported 4.7 ml/min left lung parenchymal bronchial blood flow. Anastomotic unilateral bronchopulmonary blood flow has been measured to be 4-9 ml/min during TUPAO and cardiopulmonary bypass, and 2-5 ml/min in isolated left lower lobes (Williams and Towbin, 1955; Parker and Smith, 1957; State et al., 1957; Bjork and McNeil, 1977; ModeU et aL, 1981). Bronchial blood flow to the left lung was not measured before occlusion in the present experiments because of potential contamination with mixed venous microspheres in the unoccluded state. Bjork and McNeil (1977) reported that right bronchial blood flow in rabbits was 30 % greater than that of the left which corresponds to the size differential of the two lungs, and suggests that flow/g should be equal in the two lungs. Bronchial blood flow to the right lung in the present study was substantially higher than that to the left lung (fig. 1). Apart from artifactitious technical reasons for right bronchial blood flows being substantially greater than left bronchial blood flows, TUPAO may have produced an increase in left bronchial vascular resistance as has been shown to occur by Malik and Tracy (1980) 1 h following unilateral micro-embolization. Thus, it would seem that unilateral pulmonary arterial occlusion either reduced or did not change ipsilateral bronchial blood flow; however, the degree of this apparent reduction in unilateral bronchial blood flow is masked by spillage of untrapped microspheres from systemic organs into the contralateral pulmonary artery. An increase in bronchial blood flow following TUPAO also cannot be excluded. Long term (i.e., days to months) TUPAO is known to cause an increase in bronchial blood flow. Thus, the effects of short term (i.e., rain to h) TUPAO per se remain unknown. The effects of 10 cm 1-120 P E E P on bronchial blood flow. Whether or not unilateral pulmonary arterial occlusion per se significantly altered bronchial blood flow, the
addition of 10 cm H20 PEEP which raised mean airway pressure from 8.3 cm H20 to 18.5 cm H20 produced a 3-4-fold increase in left bronchial arterial resistance. This increased resistance coupled with reduced driving pressure caused the left bronchial blood flow to fall by 70-80%. Passage of a similar amount of time without PEEP did not cause a diminution bronchial blood flow. Bruner and Schmidt (1947) also demonstrated nearly complete cessation of bronchial flow when mean airway pressure was raised to a level between 9 and 24 mm Hg for an unstated length of time but probably briefly. Horisberger and Rodbard (1960) had similar results for l-rain elevations in airway pressure. The increase in bronchial arterial resistance that was observed in response to ventilation with PEEP may have been a consequence of a mechanical alteration in the vessels, such as compression or longitudinal stretching, but it is possible that the increase in resistance was caused by active vasoconstriction that could have been mediated either by locally released humoral mediators or by neural reflexes. Observing the bronchial blood flow after removal of PEEP may be able to clarify this question, but these measurements were not made.
BRONCHIAL BLOOD FLOW DURING PEEP
275
Several investigators (Bruner and Schmidt, 1947; State et al., 1957; Horisberger and Robard, 1960; Lung et aL, 1976) have demonstrated vasoconstriction of the bronchial vessels in response to ct-adrenergic stimulation whether by injection of a-adrenalin or by sympathetic nerve stimulation which indicates that bronchial arterial smooth muscle has 0t-adrenergic receptors. Robertson etal. (1978) has reported that serum eatecholamines are elevated following institution of PEEP creating the appropriate milieu for ~t-adrenergic mediated vasoconstriction. Kaihara et ai. (1969) demonstrated an increase in bronchial resistance and a reduction in bronchial blood flow during a severe hemorrhage which would be consistent with 0t-adrenergic mediated vasoconstriction which occurs during hemorrhagic shock, but there are many other vasoactive substances circulating during hemorrhagic shock that also could be implicated. Nonetheless, the study by Kaihara etaL (1969) demonstrates that the bronchial circulation responds to an altered pathophysiological state, as well as to administration of pharmacologic mediators, by actively vasoconstricting. Thus, the reduction in bronchial blood flow with PEEP, as well as that seen with hemorrhage, may represent ct-adrenergic mediated vasoconstriction that is initiated by hypotension and reduction in baroreceptor stimulation. The effects of alterations in acid-base status on bronchial blood flow have not been determined, but the mild systemic metabolic acidosis that developed during PEEP probably has little effect compared to the local tissue alkalosis that would be expected to occur in the tissue of a ventilated lung that had virtually no CO: delivered to it. A severe tissue alkalosis would be expected to increase vascular resistance and could have contributed to the reduction in bronchial blood flow. Bruner and Schmidt's studies (1947) showing a reduction in bronchial blood flow consequent to increasing airway pressure, were carried out for an unstated but probably brief (1-2 min) period oftime. Our present data indicate that, whatever the mechanisms may be, the reduction in bronchial arterial blood flow persisted at least 2 h. Brief (min) reductions in bronchial blood flow would be expected to impose no serious harm, but the fact that the reduction in bronchial blood flow following PEEP was extreme and persisted up to 2 h in a lung void of pulmonary artery blood flow indicates that delivery of nutrients and oxygen and the removal of metabolic products in this setting may be inadequate to sustain normal cellular function of the lung. Significance of PEEP during TUPAO. Regional pulmonary arterial obstruction is thought to occur in association with high permeability pulmonary edema caused either by aggregation of cellular components (Malik and Tracy, 1980) or fibrin (Saldeen, 1976). Clinically, these are the situations in which PEEP is utilized. In addition, during ventilation with PEEP, pulmonary capillaries may become occluded because of high alveolar pressures. This would be particularly likely to occur in regions of lung that were relatively uninvolved in the original pathologic process because regions with consolidation are relatively protected (in comparison to well aerated regions) from the influence of changes in airway pressure. Thus, there are several reasons to consider that pulmonary arterial occlusion may be a complicating feature in the clinical setting that utilizes ventilation with PEEP.
276
S.C. CASSIDY AND M.S. HAYNES
Pulmonary artery occlusion alone is not considered to be sufficient ischemia to cause necrosis of lung tissue if the bronchial circulation is intact (Ellis et al., 1951; Williams and Towbin, 1955). In recent studies (Johnson et al., 1981) we have demonstrated with physiologic and histologic techniques that 10 cm H:O PEEP increased pulmonary edema and hemorrhage three-fold distal to unilateral p-lmonary arterial occlusion compared to TUPAO without PEEP; thus, evidence exists that PEEP enhances injury to the lung following pulmonary arterial occlusion. The present data substantiate that one potential mechanism for this PEEP-related injury could be a PEEP-induced reduction in bronchial blood flow.
Appendix Distribution ofraicrospheres to mixed venous blood. The appearance of microspheres in mixed venous blood was documented for 3 rain. Before PEEP was applied, the time interval during which the largest number ofmicrospheres appeared in mixed venous blood was within the first 30 sec (fig. 3). After PEEP was applied the peak appearance of microspheres in mixed venous blood was delayed to the 30-60 sec interval. This contamination seems small, rarely exceeding 50 counts per vial above background (each vial contained 5 ml), in comparison to background counts that ranged between 100 and 250 counts per empty vial. By 180 sec the leakage averaged less than 5 counts per vial above background. This seemingly small amount of leakage would amount to 2000 counts/rain assuming a 2 L cardiac output. Even a leak of microspheres into mixed venous blood at this slow rate, over a period of 2 h could add up to a substantial collection of microspheres in the lung if leakage were to continue even at a rate that is below our ability to detect. Two thousand counts/rain would add up to 240000 counts over a 2-h period, and 240000 counts, when trapped in the lung, would yield a calculated blood flow of 115 ml/min in the present experiments (averaged nuclide reference factor was 0.00048 i 2 ml/count). Since average right lung weight was 94 g, the flow/g, which reflects
160 t40
• Pre Peep 1/:) hr Followln9 10 cm HzO Peep A 1 hr Following 10 em H20 Peep o 2 hr Following 10 cm H20 Peep
~120 u
J
•:( 1oo
i.o § oo
./o
j
o 0-30
30-60
60-90 90~1Z0 120-150 SAMPLING INTERVAL (sec)
t50-180
Fig. 3. The microsphere counts appearing in mixed venous blood during the various 30 sec sampling intervals. Each symbol represents the mean value obtained from 4-6 dogs. During early sampling intervals, roughly 25 counts appeared per 5-ml vial; whereas, by 3 rain the leakage rate was reduced to 5 counts per 5-ml vial. This is below the sensitivity of the method.
277
BRONCHIAL BLOOD FLOW DURING PEEP
8m 5 o
4
2
.
.
|
• I
i
-
-
~/////II////////////////////~ ~0 em Hzo PEeP ~II//////////I/////////////////M
0
0.5
1.0 TIME (hr)
2.0
Fig. 4. The leakage ofmicrospheres into mixed venous blood expressed as a ratio of the total mixed venous counts sampled over 180 sec to the total arterial counts sampled over 90 sec. Both sources are sampled at 10 ml/min. Each symbol represents the ratio of mixed venous/arterial counts of one microsphere injection in one dog. The fraction of microspheres that leaked into mixed venous blood ranged from approximately 0.1 to 1.5% averaging 0.5%, which represents the minimum leak that could have occurred in these experiments. the maximum error in estimating right bronchial blood that could be caused by leakage of microspheres would be 1.23 ml. rain - i. g - ~. Since this was three times the actual right bronchial blood flow, we know that microspheres did not continue to leak at a rate that would yield a detectable number of microspheres. The quantity of leak which occurred could not be greater than the total microsphere count that was measured. As illustrated in fig. 4 the fraction ofmicrospheres which appeared in mixed venous blood within the 3-rain period ranged from 0.1 to 1.5% (averaging 0.5%). A 0.5~ leak would represent a blood flow of 10-15 ml/min in animals of the size used in these experiments with anticipated cardiac output of 2-3 L/rain. The average right lung blood flow (RUL and RLL in ml/min ranged from 42.3 ml/min without PEEP to 10.8 ml/min after 2 h of ventilation with 10 cm H20 PEEP. Thus, a variable small leak (0.5%) could have influenced the bronchial blood flow significantly if the pulmonary artery to the region being studied were not occluded. in this preparation blood flows estimated by microsphere dispersion to the left lung unequivocally represented left bronchial artery blood flow, and to the right lung represented right bronchial artery blood flow plus an unknown fraction of microspheres that were not trapped on one pass through the systemic organs. Data from both lungs were presented, but our conclusions are based only on data derived from the leR lung.
Acknowledgements. Our sincere appreciation is extended to Christine Smith for preparation of this manuscript and to Usha Rajagopal and Michael Pagel for their technical assistance and aid in processing the data.
References Archie, J. P., D.E. Fixler, D.J. Ullyot, J. I. E. Hoffman, J. R. Utley and E. L, Carlson (1973). Measurements of cardiac output with and organ trapping of microspheres. J. Appl. Physiol. 35: 148-154.
278
S.C. CASSIDY AND M.S. HAYNES
Baile, E. M., J. M. B. Nelems, M. Schulzer and P. D. Pare (1982). Measurement of regional bronchial arterial blood flow and broncbovascular resistance in dogs. J. Appl. Physiol. 53: 1044-1049. Bjork, L. and B.J. McNeil (1977). Blood flow in pulmonary and bronchial arteries in acute experimental pneumonia and pulmonary embolism. Acta Radiol. Diagn. 18: 393-399. Bruner, H.D. and C.F. Schmidt (1947). Blood flow in the bronchial artery of the anesthesized dog. Am. J. Physiol. 148: 648-666. Domenech, R.J., J. I. E. Hoffman, M. I.M. Nobel, K.B. Saunders, J.R. Henson and S. Subijanto (1969). Total and regional coronary blood flow measured by radioactive microspberes in conscious and anesthetized dogs. Circ. Res. 25: 581-596. Ellis, F. H., Jr., J. H. Grindlay and J. E. Edwards (1951). The bronchial arteries: I. Experimental occlusion. Surgery 30: 810-826. Haynes, M. S., S. S. Cassidy and R. L. Johnson, Jr. (1979). The effect of PEEP on bronchial blood flow. Fed. Proc. 38: 1378. Horisberger, B. and S. Rodbard (1960). Direct measurement of bronchial arterial flow. Circ. Res. 8: ! 149-1156. Johnson, R. L., Jr., S. S. Cassidy, M. S. Haynes, R. L. Reynolds and W. Schultz (1981). Microvascular injury distal to unilateral pulmonary artery occlusion. J. Appl. Physiol. 51: 845-851. Kalhara, S., R. B. Rutherford, E. P. Schwentker and H. N. Wagner, Jr. (1969). Distribution of cardiac output in experimental hemorrhagic shock in dogs. J. Appl. Physiol. 27: 218-222. Lung, M. A. K.Y., J. C. C. Wang and K.K. Cheng (1976). Bronchial circulation: An auto-perfusion method for assessing its vasomotor activity and the study of alpha- and beta-adrenoceptors in the bronchial artery. L~fe Sci. 19: 577-580. Malik, A.B. and S.E. Tracy (1980). Bronchovascular adjustments al~er pulmonary embolism. J. Appl. Physiol. 49: 476-481. Modell, H.I., K. Beck and J. Butler (1981). Functional aspects of canine bronchial-pulmonary vascular communication. J. Appl. Physiol. 50:1045-1051. Parker, B.M. and J.R. Smith (1957). Studies of experimental pulmonary embolism and infarction and development of collateral circulation in the affected lung lobe. J. Lab. Clin. Med. 49: 850-857. Pietra, G.G. and M. Magno (1978). Pharmacological factors influencing permeability of the bronchial microcirculation. Fed. Proc. 37: 2466-2470. Robertson, C. H., Jr., A.C. Weihl and M. E. Bradley (1978). Plasma catechol changes of intermittent positive pressure breathing with positive end-expiratory pressure. Am. Rev. Reypir. DL~. 117: 385. Rudolph, A.M. and M.A. Heymann (1967). The circulation of the fetus in utero; methods for studying distribution of blood flow, cardiac output and organ blood flow. Circ. Res. 21: 163-184. Saldeen, T. (1976). Microembolization syndrome. Microvasc. Res. 11 : 227-259. State, D., P. F. Salisbury and P. Well (1957). Physiologic and pharmacologic studies of collateral pulmonary flow. J. Thoracic Surg. 34: 599-608. Williams, M. H., Jr., and E.J. Towbin (1955). Magnitude of time of development of the collateral circulation to the lung after occlusion of the lea pulmonary artery. Circ. Res. 3: 422-424. Zar, J.H. (1974). Biostatistical Analysis. Englewood Cliffs, NJ, Prentice Hall, Inc.
Respiration Physiology (1986) 66, 269-278 Elsevier
THE EFFECTS OF VENTILATION WITH POSITIVE END-EXPIRATORY PRESSURE ON THE BRONCHIAL CIRCULATION*
SHARON S. CASSIDY and MICHAEL S. HAYNES Laboratoryfor Cardiopulmonary Research and the Pulmonary Research Division of the Department of lnternal Medicine, University of Texas Southwestern Medical School Dallas, 75233, U.S.A.
Abstract. We studied the effects of ventilation with 10 cm H20 PEEP for 2 h in dogs with temporary unilateral pulmonary arterial occlusion (TUPAO) on bronchial blood flow to the occluded lung using the microsphere dispersion technique. We found that blood flow to the occluded left lung in dogs was 9.9 ml/min ( 0 . 1 2 2 m l ' m i n - t ' g - I ) . Within 30min following the addition of 10cm H20 PEEP blood flow fell by 70-80 ~ (to 2.3 ml/min) caused both by a 3-fold decrease in vascular conductance and a 25 % fall in systemic blood pressure. The reduction in left bronchial blood flow persisted for at least 2 h. We conclude from these data that ventilation with PEEP in the presence of pulmonary artery occlusion has a severe, persistent adverse effect on bronchial blood flow. This reduction in bronchial blood flow is beyond what can be explained by the changes in airway pressure. The additional increase in bronchial vascular resistance may be caused by the increase in lung volume, by reflex bronchial vasoconstriction, or by release of mediators locally. Dog Lung injury
Microsphere PEEP
TUPAO
Aside from its critical role in various types of congenital heart disease, bronchial circulation has been implicated as being potentially important in the formation of pulmonary edema caused by increased permeability (Pietra and Magno, 1978) and in the sequelae of pulmonary artery occlusion (Ellis et al., 1951; William and Towbin, 1955; Parker and Smith, 1957). Johnson et al. (1981) recently studied the effects of a temporary unilateral pulmonary arterial occlusion (TUPAO) on a group of dogs and found that by histological assessment as well as by measurement of lung tissue volume (C2H 2 method), the lung distal to the TUPAO developed edema and injury. The extent of the injury was enhanced by ventilation with 10 cm H20 PEEP. We reasoned that Accepted for publication 22 August 1986 * A preliminary report was presented to the Federation of American Societies of Experimental Biology, April, 1979 in Dallas, Texas (Haynes etal., 1979) 0034-5687/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)
270
S.C. CASSIDY AND M.S. HAYNES
an adverse effect of PEEP on bronchial blood flow in the setting of TUPAO could have contributed to lung injury, and the purpose of the present study was to determine the effects of PEEP on bronchial arterial blood flow during TUPAO using the microsphere dispersion technique. The microsphere method for measuring organ blood flow has not been applied often to the bronchial circulation because of the potential leakage of microspheres from the systemic circulation into venous blood with subsequent entrapment in the lungs via the pulmonary circulation. Because pulmonary arterial blood flow to one lung was occluded in this study, the potential for contamination with microspheres which leaked through the systemic circulation was eliminated for the occluded lung.
Methods
Twelve mongrel dogs of either sex weighing 18-22 kg were studied under pentobarbital anesthesia, 25 to 30 ml/kg, i.v. They were ventilated with a large animal sinusoidal pump ventilator at 15 ml/kg via an endotracheal tube at a rate sufficient to maintain medal Pco2 at approximately 40 mm Hg. During left thoracotomy a catheter was placed in the left atrial appendage for subsequent injection of radionuclide microspheres. The chest was closed tightly and residual air was aspirated from the pleural space. Two catheters were placed in the femoral artery, one for sampling during microsphere injections, the other for monitoring medal pressure. In addition, catheters were advanced to the right atrium and pulmonary artery via femoral veins. A balloon-occlusion catheter was placed in the left pulmonary artery and was inflated to occlude blood flow. Thirty minutes following occlusion, measurements of blood flow distributions were made without PEEP (IPPV). In nine dogs additional measurements were made 30, 60, and 120 min following the addition of 10 cm H20 PEEP. In 3 dogs measurements were made at similar intervals but without adding PEEP. Arterial blood gases, pH, and pressure recordings were obtained prior to each blood flow measurement. The technique for measuring organ blood flow using dispersion of radionuclide microspheres was adapted from the original description of Rudolph and Heymann (1967) using the reference sampling adaption ofArchie et al. (1973) and Domenech et al. (1969). Radionuclides used in random order were 14ICe, 125I, SSSr, and 46Sc. A 3-ml mixture containing 500 000 25-/~m latex spheres suspended in Tween was injected via the left atrial catheter and flushed with 30 ml warmed saline over 30 sec. The time to clear the dead space of the blood withdrawal system which pumped at 10 ml/min was 20 sec and was independent of medal pressure. After a 20 sec delay, blood was also sampled from the pulmonary artery for 180 sec to detect microsphere spillover from the systemic vascular bed. At the termination of the experiment, the animals were killed by exsanguination, organs were removed, and connective tissue was dissected away. The lungs were divided into right and left, upper and lower lobes, and these tissues were minced and placed in preweighed vials. The vials containing tissue and blood were
BRONCHIAL BLOOD FLOW DURING PEEP
271
reweighed and placed in a scintillation counter for determining the gamma emission at preselected channels corresponding to the peak KEV emission of each nuclide. Spillover of emission from one nuclide into channels detecting the peak KEV of the other nuclides was subtracted according to the methods previously described by Rudolph and Heymann (1967).
Statistical analyses.
Data are expressed as means + SEM. Data were compared with a one way analysis of variance for repeated measures (Zar, 1974). When the F value indicated a significant non-random difference (P < 0.05) among repeated measures, Newman-Keuls multiple comparison procedure was used to determine significance of difference between pairs of means (Zar, 1974). P < 0.05 was chosen as the level reflecting significance.
Results
Bronchial arterial blood flow during unilateral pulmonary artery occlusion and PEEP (fig. 1). Following left TUPAO bronchial blood flow measured 9.9 ml/min. Blood flow was higher to the left lower lobe compared to the left upper lobe, 0.135 vs 0.110 ml. rain - t .g - ,, and this trend continued following PEEP. Blood flow to the left lung (with the left pulmonary artery occluded) fell by 70% to 72% within 30 min after placing the animal on 10 cm H20 PEEP. Through the next 1.5 h bronchial blood flow continued to fall until it was 84-85% below pre-PEEP levels. Blood flow to the occluded left lung in three dogs that were not placed on PEEP did not change significantly (falling by 20%) over a period of 1.5 h following the fu'st measurement of blood flow 30 rain post occlusion.
Bronchial arterial driving pressure during unilateral pulmonary arterial occlusion and PEEP (fig. I). The driving pressure for bronchial blood flow should be the difference between systemic m e d a l pressure and either pulmonary venous or alveolar pressure depending o • o • I
Ls.
RIght to',,,m"k~:~e Right t#ppe¢Iot~ Left I o ~ r lobe Left ~oper I o ~ tS.E.
o
S~em~c orfery
3OO
(SA)
• .~A-wq~dge
E 2
100 qcz
0 ~e 30 PEEP
60
90
t20
I Pre
1 30
I 60
I 90
I 120
0
PEEP TIME DURING tOcm H20 PEEP (rain)
Fig. 1. Bronchialarterial blood flows and bronchial arterial driving pressures followingleft pulmonary arterial occlusionbefore and during2 h ventilationwith 10 cm H20 PEEP.
272
s.c. CASSIDY AND M.S. HAYNES
TABLE 1 Pressures that may influence bronchial blood flowduring temporary unilateral pulmonaryartery occlusion before and durillg ventilation with 10 ¢m H20 PEEP. Pulmonary arterial pressure (cm H20)
Pulmonary wedge pressure (cm H20)
Systemic arterial pressure (mm Hg)
Airway pressure (cm H20)
Before PEEP
28.3 _+2.5
9.6 4. 1.7
148 4- 13
8.3 _+0.9
During PEEP 30 min
28.6 _4-5.5
13.3" + 2.5
88* 4. 8
16.4" _+1.4
60 rain
30.6 4_4.6
13.3 4_2.9
102 +_11
18.1 +0.8
120 min
28.9 +3.7
11.9 +3.1
95 4. 7
18.6 4.1.0
Values are means + SEM, n = 6. * P < 0.05 pre-PEEP vs 30 rain post-PEEP. All comparisons between 30 min and 120 rain were not significant.
on which system (pulmonary venous or alveolar) has the highest critical closing pressure to bronchial outflow. Absolute pulmonary arterial, pulmonary wedge, systemic arterial and airway pressures are listed in table 1. Airway pressures rose from 8 cm H 2 0 to 16.4-18.6 cm H 2 0 with PEEP and pulmonary arterial mean pressure did not change significantly. Depending on which measure was used, driving pressure fell either by 49% or 46% initially with 10 cm H 2 0 PEEP. With further passage of time driving pressure did not continue to fall but was restored slightly ( 3 - 6 % ) by virtue of the restoration of systemic arterial pressure.
Left bronchial vascular resistance during TUPA O and PEEP (fig. 2).
Bronchial vascular resistance to the occluded left lung increased 2-3 fold with the addition of 10 cm H 2 0 PEEP. With passage of time left bronchial vascular resistance rose further so that after 2 h the resistance was 4 times greater than pre-PEEP values. The conductance of the left bronchial vascular channels fell to 1/2 to 1/3 of pre-PEEP levels after 2 h of ventilation with PEEP.
Distribution of microspheres to mixed venous blood. See Appendix. Arterial and venous pH and blood gases (table 2).
A mild metabolic acidosis developed over the 2 h period during T U P A O and PEEP. Pao2 remained above 100 mm Hg while mixed venous Po, fell and Pco2 increased both probably reflecting a low output state.
BRONCHIAL BLOOD FLOW DURING PEEP
273
g
'c E "F i
Ns
1.0
¢
-,..i. . . . . .I. . . . .
E E
i
Ns
5"
os 33
o ~Z
3
i
== I
I
I
!
I
l
I
I
0
30
6O
120
0
3O
6O
f20
0
I
Time After 10 cm H20 PEEP (rain)
Fig. 2. Left bronchial arterial resistance and conductance followingleft TUPAO before and during 2 h ventilation with 10 cm H20 PEEP. TABLE 2 Arterial and mixed venous blood gas tensions and pH
pH
Before PEEP
Pco2
Po2
Arterial
Mixed venous
Arterial
Mixed venous
Arterial
Mixed venous
7.41 ± 0.03
7.39 _+0.02
33.7 _+2.5
38.3 ± 3.5
100 _+ 5
39 ±3
7.29* _+0.04
30.3 _+6.7
43.0* _+8.2
113 ± 13
32* ±7
Following 10 cm H20 PEEP 30 rain 7.35* _+0.05 60 rain
7.33 ± 0.02
7.28 ± 0.02
32.0 ± 6.9
41.7 _+7.6
113 _+12
36 4-4
120 min
7.301. _+0.02
7.231" _+0.01
34.0 _+5.2
51.3' _+8.1
117 _+23
291" _+4
Means± SE; * P < 0.05 pre-PEEP vs 30 min post-PEEP; 1.P < 0.05 post-PEEP 30 min vs 120 min. L u n g weights. Averaged post m o r t e m wrights of the lung were: left upper lobe, 31.8 +_ 1.9g; left lower lobe, 47.5 _+ 2 . 7 g ; left lung 81.0 _+ 4 . 7 g ; fight upper lobe, 42.7 _+ 1.89 g; fight lower lobe, 52.1 +_ 3 . 0 g ; a n d right lung, 94.8 +_ 4.2g.
Discussion Bronchial blood flow to the entire left lung averaged 9.9 ml/min. This corresponds reasonably with m e a s u r e m e n t s made with flow meters in lungs receiving their n o r m a l
274
s.c. CASSIDY AND M.S. HAYNES
complement of pulmonary arterial flow. Bruner and Schmidt (1947) reported right posterior bronchial blood flow (60~ of all right bronchial flow) was 4.8 ml/min, Horisberger and Rodbard (1960) reported bilateral bronchial blood flows of 12.1 ml/min, and Baile et al. (1982) reported 4.7 ml/min left lung parenchymal bronchial blood flow. Anastomotic unilateral bronchopulmonary blood flow has been measured to be 4-9 ml/min during TUPAO and cardiopulmonary bypass, and 2-5 ml/min in isolated left lower lobes (Williams and Towbin, 1955; Parker and Smith, 1957; State et al., 1957; Bjork and McNeil, 1977; ModeU et aL, 1981). Bronchial blood flow to the left lung was not measured before occlusion in the present experiments because of potential contamination with mixed venous microspheres in the unoccluded state. Bjork and McNeil (1977) reported that right bronchial blood flow in rabbits was 30 % greater than that of the left which corresponds to the size differential of the two lungs, and suggests that flow/g should be equal in the two lungs. Bronchial blood flow to the right lung in the present study was substantially higher than that to the left lung (fig. 1). Apart from artifactitious technical reasons for right bronchial blood flows being substantially greater than left bronchial blood flows, TUPAO may have produced an increase in left bronchial vascular resistance as has been shown to occur by Malik and Tracy (1980) 1 h following unilateral micro-embolization. Thus, it would seem that unilateral pulmonary arterial occlusion either reduced or did not change ipsilateral bronchial blood flow; however, the degree of this apparent reduction in unilateral bronchial blood flow is masked by spillage of untrapped microspheres from systemic organs into the contralateral pulmonary artery. An increase in bronchial blood flow following TUPAO also cannot be excluded. Long term (i.e., days to months) TUPAO is known to cause an increase in bronchial blood flow. Thus, the effects of short term (i.e., rain to h) TUPAO per se remain unknown. The effects of 10 cm 1-120 P E E P on bronchial blood flow. Whether or not unilateral pulmonary arterial occlusion per se significantly altered bronchial blood flow, the
addition of 10 cm H20 PEEP which raised mean airway pressure from 8.3 cm H20 to 18.5 cm H20 produced a 3-4-fold increase in left bronchial arterial resistance. This increased resistance coupled with reduced driving pressure caused the left bronchial blood flow to fall by 70-80%. Passage of a similar amount of time without PEEP did not cause a diminution bronchial blood flow. Bruner and Schmidt (1947) also demonstrated nearly complete cessation of bronchial flow when mean airway pressure was raised to a level between 9 and 24 mm Hg for an unstated length of time but probably briefly. Horisberger and Rodbard (1960) had similar results for l-rain elevations in airway pressure. The increase in bronchial arterial resistance that was observed in response to ventilation with PEEP may have been a consequence of a mechanical alteration in the vessels, such as compression or longitudinal stretching, but it is possible that the increase in resistance was caused by active vasoconstriction that could have been mediated either by locally released humoral mediators or by neural reflexes. Observing the bronchial blood flow after removal of PEEP may be able to clarify this question, but these measurements were not made.
BRONCHIAL BLOOD FLOW DURING PEEP
275
Several investigators (Bruner and Schmidt, 1947; State et al., 1957; Horisberger and Robard, 1960; Lung et aL, 1976) have demonstrated vasoconstriction of the bronchial vessels in response to ct-adrenergic stimulation whether by injection of a-adrenalin or by sympathetic nerve stimulation which indicates that bronchial arterial smooth muscle has 0t-adrenergic receptors. Robertson etal. (1978) has reported that serum eatecholamines are elevated following institution of PEEP creating the appropriate milieu for ~t-adrenergic mediated vasoconstriction. Kaihara et ai. (1969) demonstrated an increase in bronchial resistance and a reduction in bronchial blood flow during a severe hemorrhage which would be consistent with 0t-adrenergic mediated vasoconstriction which occurs during hemorrhagic shock, but there are many other vasoactive substances circulating during hemorrhagic shock that also could be implicated. Nonetheless, the study by Kaihara etaL (1969) demonstrates that the bronchial circulation responds to an altered pathophysiological state, as well as to administration of pharmacologic mediators, by actively vasoconstricting. Thus, the reduction in bronchial blood flow with PEEP, as well as that seen with hemorrhage, may represent ct-adrenergic mediated vasoconstriction that is initiated by hypotension and reduction in baroreceptor stimulation. The effects of alterations in acid-base status on bronchial blood flow have not been determined, but the mild systemic metabolic acidosis that developed during PEEP probably has little effect compared to the local tissue alkalosis that would be expected to occur in the tissue of a ventilated lung that had virtually no CO: delivered to it. A severe tissue alkalosis would be expected to increase vascular resistance and could have contributed to the reduction in bronchial blood flow. Bruner and Schmidt's studies (1947) showing a reduction in bronchial blood flow consequent to increasing airway pressure, were carried out for an unstated but probably brief (1-2 min) period oftime. Our present data indicate that, whatever the mechanisms may be, the reduction in bronchial arterial blood flow persisted at least 2 h. Brief (min) reductions in bronchial blood flow would be expected to impose no serious harm, but the fact that the reduction in bronchial blood flow following PEEP was extreme and persisted up to 2 h in a lung void of pulmonary artery blood flow indicates that delivery of nutrients and oxygen and the removal of metabolic products in this setting may be inadequate to sustain normal cellular function of the lung. Significance of PEEP during TUPAO. Regional pulmonary arterial obstruction is thought to occur in association with high permeability pulmonary edema caused either by aggregation of cellular components (Malik and Tracy, 1980) or fibrin (Saldeen, 1976). Clinically, these are the situations in which PEEP is utilized. In addition, during ventilation with PEEP, pulmonary capillaries may become occluded because of high alveolar pressures. This would be particularly likely to occur in regions of lung that were relatively uninvolved in the original pathologic process because regions with consolidation are relatively protected (in comparison to well aerated regions) from the influence of changes in airway pressure. Thus, there are several reasons to consider that pulmonary arterial occlusion may be a complicating feature in the clinical setting that utilizes ventilation with PEEP.
276
S.C. CASSIDY AND M.S. HAYNES
Pulmonary artery occlusion alone is not considered to be sufficient ischemia to cause necrosis of lung tissue if the bronchial circulation is intact (Ellis et al., 1951; Williams and Towbin, 1955). In recent studies (Johnson et al., 1981) we have demonstrated with physiologic and histologic techniques that 10 cm H:O PEEP increased pulmonary edema and hemorrhage three-fold distal to unilateral p-lmonary arterial occlusion compared to TUPAO without PEEP; thus, evidence exists that PEEP enhances injury to the lung following pulmonary arterial occlusion. The present data substantiate that one potential mechanism for this PEEP-related injury could be a PEEP-induced reduction in bronchial blood flow.
Appendix Distribution ofraicrospheres to mixed venous blood. The appearance of microspheres in mixed venous blood was documented for 3 rain. Before PEEP was applied, the time interval during which the largest number ofmicrospheres appeared in mixed venous blood was within the first 30 sec (fig. 3). After PEEP was applied the peak appearance of microspheres in mixed venous blood was delayed to the 30-60 sec interval. This contamination seems small, rarely exceeding 50 counts per vial above background (each vial contained 5 ml), in comparison to background counts that ranged between 100 and 250 counts per empty vial. By 180 sec the leakage averaged less than 5 counts per vial above background. This seemingly small amount of leakage would amount to 2000 counts/rain assuming a 2 L cardiac output. Even a leak of microspheres into mixed venous blood at this slow rate, over a period of 2 h could add up to a substantial collection of microspheres in the lung if leakage were to continue even at a rate that is below our ability to detect. Two thousand counts/rain would add up to 240000 counts over a 2-h period, and 240000 counts, when trapped in the lung, would yield a calculated blood flow of 115 ml/min in the present experiments (averaged nuclide reference factor was 0.00048 i 2 ml/count). Since average right lung weight was 94 g, the flow/g, which reflects
160 t40
• Pre Peep 1/:) hr Followln9 10 cm HzO Peep A 1 hr Following 10 em H20 Peep o 2 hr Following 10 cm H20 Peep
~120 u
J
•:( 1oo
i.o § oo
./o
j
o 0-30
30-60
60-90 90~1Z0 120-150 SAMPLING INTERVAL (sec)
t50-180
Fig. 3. The microsphere counts appearing in mixed venous blood during the various 30 sec sampling intervals. Each symbol represents the mean value obtained from 4-6 dogs. During early sampling intervals, roughly 25 counts appeared per 5-ml vial; whereas, by 3 rain the leakage rate was reduced to 5 counts per 5-ml vial. This is below the sensitivity of the method.
277
BRONCHIAL BLOOD FLOW DURING PEEP
8m 5 o
4
2
.
.
|
• I
i
-
-
~/////II////////////////////~ ~0 em Hzo PEeP ~II//////////I/////////////////M
0
0.5
1.0 TIME (hr)
2.0
Fig. 4. The leakage ofmicrospheres into mixed venous blood expressed as a ratio of the total mixed venous counts sampled over 180 sec to the total arterial counts sampled over 90 sec. Both sources are sampled at 10 ml/min. Each symbol represents the ratio of mixed venous/arterial counts of one microsphere injection in one dog. The fraction of microspheres that leaked into mixed venous blood ranged from approximately 0.1 to 1.5% averaging 0.5%, which represents the minimum leak that could have occurred in these experiments. the maximum error in estimating right bronchial blood that could be caused by leakage of microspheres would be 1.23 ml. rain - i. g - ~. Since this was three times the actual right bronchial blood flow, we know that microspheres did not continue to leak at a rate that would yield a detectable number of microspheres. The quantity of leak which occurred could not be greater than the total microsphere count that was measured. As illustrated in fig. 4 the fraction ofmicrospheres which appeared in mixed venous blood within the 3-rain period ranged from 0.1 to 1.5% (averaging 0.5%). A 0.5~ leak would represent a blood flow of 10-15 ml/min in animals of the size used in these experiments with anticipated cardiac output of 2-3 L/rain. The average right lung blood flow (RUL and RLL in ml/min ranged from 42.3 ml/min without PEEP to 10.8 ml/min after 2 h of ventilation with 10 cm H20 PEEP. Thus, a variable small leak (0.5%) could have influenced the bronchial blood flow significantly if the pulmonary artery to the region being studied were not occluded. in this preparation blood flows estimated by microsphere dispersion to the left lung unequivocally represented left bronchial artery blood flow, and to the right lung represented right bronchial artery blood flow plus an unknown fraction of microspheres that were not trapped on one pass through the systemic organs. Data from both lungs were presented, but our conclusions are based only on data derived from the leR lung.
Acknowledgements. Our sincere appreciation is extended to Christine Smith for preparation of this manuscript and to Usha Rajagopal and Michael Pagel for their technical assistance and aid in processing the data.
References Archie, J. P., D.E. Fixler, D.J. Ullyot, J. I. E. Hoffman, J. R. Utley and E. L, Carlson (1973). Measurements of cardiac output with and organ trapping of microspheres. J. Appl. Physiol. 35: 148-154.
278
S.C. CASSIDY AND M.S. HAYNES
Baile, E. M., J. M. B. Nelems, M. Schulzer and P. D. Pare (1982). Measurement of regional bronchial arterial blood flow and broncbovascular resistance in dogs. J. Appl. Physiol. 53: 1044-1049. Bjork, L. and B.J. McNeil (1977). Blood flow in pulmonary and bronchial arteries in acute experimental pneumonia and pulmonary embolism. Acta Radiol. Diagn. 18: 393-399. Bruner, H.D. and C.F. Schmidt (1947). Blood flow in the bronchial artery of the anesthesized dog. Am. J. Physiol. 148: 648-666. Domenech, R.J., J. I. E. Hoffman, M. I.M. Nobel, K.B. Saunders, J.R. Henson and S. Subijanto (1969). Total and regional coronary blood flow measured by radioactive microspberes in conscious and anesthetized dogs. Circ. Res. 25: 581-596. Ellis, F. H., Jr., J. H. Grindlay and J. E. Edwards (1951). The bronchial arteries: I. Experimental occlusion. Surgery 30: 810-826. Haynes, M. S., S. S. Cassidy and R. L. Johnson, Jr. (1979). The effect of PEEP on bronchial blood flow. Fed. Proc. 38: 1378. Horisberger, B. and S. Rodbard (1960). Direct measurement of bronchial arterial flow. Circ. Res. 8: ! 149-1156. Johnson, R. L., Jr., S. S. Cassidy, M. S. Haynes, R. L. Reynolds and W. Schultz (1981). Microvascular injury distal to unilateral pulmonary artery occlusion. J. Appl. Physiol. 51: 845-851. Kalhara, S., R. B. Rutherford, E. P. Schwentker and H. N. Wagner, Jr. (1969). Distribution of cardiac output in experimental hemorrhagic shock in dogs. J. Appl. Physiol. 27: 218-222. Lung, M. A. K.Y., J. C. C. Wang and K.K. Cheng (1976). Bronchial circulation: An auto-perfusion method for assessing its vasomotor activity and the study of alpha- and beta-adrenoceptors in the bronchial artery. L~fe Sci. 19: 577-580. Malik, A.B. and S.E. Tracy (1980). Bronchovascular adjustments al~er pulmonary embolism. J. Appl. Physiol. 49: 476-481. Modell, H.I., K. Beck and J. Butler (1981). Functional aspects of canine bronchial-pulmonary vascular communication. J. Appl. Physiol. 50:1045-1051. Parker, B.M. and J.R. Smith (1957). Studies of experimental pulmonary embolism and infarction and development of collateral circulation in the affected lung lobe. J. Lab. Clin. Med. 49: 850-857. Pietra, G.G. and M. Magno (1978). Pharmacological factors influencing permeability of the bronchial microcirculation. Fed. Proc. 37: 2466-2470. Robertson, C. H., Jr., A.C. Weihl and M. E. Bradley (1978). Plasma catechol changes of intermittent positive pressure breathing with positive end-expiratory pressure. Am. Rev. Reypir. DL~. 117: 385. Rudolph, A.M. and M.A. Heymann (1967). The circulation of the fetus in utero; methods for studying distribution of blood flow, cardiac output and organ blood flow. Circ. Res. 21: 163-184. Saldeen, T. (1976). Microembolization syndrome. Microvasc. Res. 11 : 227-259. State, D., P. F. Salisbury and P. Well (1957). Physiologic and pharmacologic studies of collateral pulmonary flow. J. Thoracic Surg. 34: 599-608. Williams, M. H., Jr., and E.J. Towbin (1955). Magnitude of time of development of the collateral circulation to the lung after occlusion of the lea pulmonary artery. Circ. Res. 3: 422-424. Zar, J.H. (1974). Biostatistical Analysis. Englewood Cliffs, NJ, Prentice Hall, Inc.