- Email: [email protected]
Temperature dependence of intraparenchymal bronchial blood flow
Respiration Elsevier
Physiology
(1987)
259
68, 259-267
RSP 01288
Temperature dependence of intraparenchymal bronchial blood flow PiergiuseppeAgostoni, Mark E. Deffebach, Wayne Kirk and George L. Brengelmann Department
of Medicine,
Division of Respiratory Diseases and University of Washington. Seattle, (Accepted
for publication
Department WA, U.S.A.
11 February
of Physiology
and
Biophysics,
1987)
Abstract. Previous studies suggested that bronchial vascular resistance, like that of the skin, changes with the temperature of the surrounding tissue. To investigate this phenomenon, we recorded anastomotic (systemic to pulmonary) (obr,.,) and total (obr) bronchial blood flow over a temperature range centered on normal. In 7 open-chested dogs the in sifu left lower lobe (LLL) was separately ventilated (30 “C, 5% CO, in humidified air) and was suspended in a fabric net from a strain gauge for continuous recording of weight. The pulmonary circulation of the LLL was pump-perfused at 255 f 69 ml/min in a closed circuit with temperature set at 30, 33.36, 39 and 42 “C. obr,., was measured as overflow from the LLL vascular circuit corrected for LLL weight changes. obr, tracheal, mid-esophageal and coronary flow were measured with 15 /J radiolabelled microspheres injected in the left atrium. The animal’s core temperature and that of the humidified air around the LLL were held constant. obr and obr,+ were equal and reached a peak at 36 “C with lower levels of flow at higher and lower temperatures. Esophageal, tracheal and coronary flow and cardiac output did not change nor did pressures in the systemic and LLL pulmonary artery and in the LLL airways. An intralobar change in temperature above or below 36 “C decreases only the lobar bronchial blood flow and does not influence blood flow to other nearby tissues including those vascularized by the bronchial circulation.
Dog;
Esophageal
blood
flow;
Left
lower
lobe;
Microspheres;
Tracheal
blood
flow
Blood flow in the bronchial vasculature appears to be strongly influenced by temperature. In a recent study of the effects of pulmonary artery occlusion (Agostoni et af., 1987) we showed that the discrepancy between measurements of intrapulmonary bronchial blood flow, both anastomotic (obr,-,) and total (Gbr), obtained in an open (Baile et al., 1982, 1985; Jindal et al., 1984) vs a closed chest preparation (Malik and Tracy, 1980) could be explained on the basis of the differing thermal conditions. The physiological significance of this temperature-dependence could stem from the fact that temperatures of mucosa of the trachea and bronchi are influenced by the external Correspondence Italy. 0034-5687/87/$03.50
address:
0
Dr.
P.G.
Agostoni,
1987 Elsevier
Science
Istituto
di Cardiologia,
Publishers
Universita
B.V. (Biomedical
di Milano,
Division)
20138
Milan,
260
P. G. AGOSTONI
ef al.
thermal environment, particularly during hyperventilation with cold air (McFadden et al., 1982). Temperature-dependence of bronchial blood flow could be due to direct thermal effects on tone in vascular smooth muscle of the large airways. A thermally mediated decrease in bronchial blood flow was suggested by the observation that the tracheal mucosa of human subjects turned pale during ventilation with cold air (McFadden, 1983). The opposite, an increase in flow with reduced temperature, has been observed (Baile et al., 1985) in the bronchial circulation of open-chested dogs hyperventilated with cold dry air ( - 20 “C). The interpretation of findings obtained with cold dry air ventilation is complicated by the likelihood that dehydration of the bronchial mucosa by the low humidity of the air may influence the airways’ muscular tone and, indirectly, the bronchial blood flow (A&ken and Marini, 1985; Baier et&, 1985). Temperature-dependent vascular tone could also be a characteristic of bronchial vessels of tissues other than the major airways. Within the pulmonary parenchyma, the influence of pulmonary arterial temperature would be dominant, but bronchial vessel temperature would in part depend upon direct thermal affects of ventilation and in part upon the temperature of arterial blood that arrives via pathways adjacent to proximal larger airways where the air may be at temperatures different from body temperature (McFadden et al., 1982). We designed our experiment to study the temperature dependence of intrapulmonary bronchial blood flow over a range of temperatures established within the parenchyma of a pump-perfused in situ left lower lobe (LLL) in an open-chested dog with the remainder of the preparation held normothermic. With this design, low lobar temperatures could be obtained without the airway drying that would have been inevitable with ventilatory heat exchange. We also used radioactive microspheres to measure flow to the trachea and the middle portion of the esophagus. In the dog, both of these are vascularized, at least partially, by the bronchial circulation (bronchoesophageal arteries; Evans and Christenen, 1979). The intent was to determine whether manipulation of temperature of the isolated lobe would affect flow in these other structures which remained at core temperature.
Methods
Our experimental design called for measurement of systemic blood flow to the trachea (otr), the esophagus toe), the heart (oh), the entire left lower lobe (obr), and the anastomotic, i.e. systemic to pulmonary bronchial blood flow (obr,-,) to the left lower lobe (LLL) with body temperature constant and with the LLL pulmonary artery blood at one of 5 selected temperatures. otr, Qe, oh and Qbr were determined with radiolabelled microspheres using a standard reference flow technique (Heymann et al., 1977) while obr,-, was measured as overflow from the isolated perfused LLL corrected for weight changes (Jindal et al., 1984). Seven mongrel dogs (24.9 + 1.7 kg, mean & SD) were anesthetized with thiopenthal
BRONCHIAL
BLOOD
FLOW
AND
LLL
TEMPERATURE
261
(30 mg/kg) followed by alpha-choralose (30 mg/kg) mixed with urethane (100 mg/kg). The LLL was isolated in situ using a technique previously described (Jindal et al., 1984). The left hemithorax was widely opened and the left upper lobe and the lingula were excised. The right lung and the LLL were ventilated separately with humidified gases (right lung 100% 0, and LLL 5% CO, in air) at a temperature of 30 “C with the end-expiratory pressure held at 5 cm H,O. LLL tidal volume was 150 cc at 12 respirations/min. Without displacing the hilar structures, the LLL was placed in a fabric net hung from a strain gauge. The LLL pulmonary circulation was isolated by inserting glass cannulas in the pulmonary artery and, via the left atrium, in the pulmonary vein. When, rarely, there was more than one vein, we cannulated the larger. We tied off the smaller and relied on the venous anastomoses for its drainage. The extracorporeal LLL vascular circuit included a reservoir with an overflow tube open to the atmosphere at the level of the lower-most portion of the LLL. This reservoir fed a roller pump (0 = 255 f 69 ml/min, mean k SD) and a coil of tubing immersed in a water bath in which the reservoir was also immersed. The LLL pump flow was slowly increased up to 350 ml/mm and/or to a LLL pulmonary artery pressure (Ppa-LLL) of 25 cm H,O. The water bath temperature was set to control the perfusate (autologous blood) temperature. To protect the dog from cooling, the thorax and abdomen were enclosed in a plexiglass box in which the air was fully humidified and maintained between 38 and 39 “C. The back of the dog was thermally isolated from the operating table with an insulating pad. We continuously recorded the LLL weight (strain gauge), the right pulmonary temperature (Trpa, Swan-Ganz), the LLL main pulmonary artery and vein temperature (thermocouples) and pressures (measured at end-expiration with zero reference at the lowermost portion of the LLL). Every 15 min, cardiac output (ot) was measured in triplicate by thermodilution. Blood gases, arterial pH and the hemodynamic status were maintained in the normal range with intravenous fluids and bicarbonate, and by adjusting the right lung ventilation. obr,-, to the LLL was measured as the volume of blood collected from the overflow tube over 5 min. To account for accumulation or loss of intra- or extra-vascular fluids in the lobe, the overflow volume was corrected for the LLL weight changes (Jindal et al., 1984). obr,-, was collected continuously for the entire experiment. obr, Qtr, oh and oe were measured with a reference flow technique (Heymann et al., 1977) by injecting approximately 2 x lo6 radiolabelled microspheres (46Sc, 95Nb, lo3Ru, “‘Sn and 14’Ce) in random order into the left atrium. The diameter of the spheres ranged from 16.5 + 0.1 pm to 15.9 + 0.5 pm. Prior to the injection the spheres were suspended by mechanical mixing and an ultrasonic bath. The reference flow was withdrawn with a constant flow pump (4.6 ml/min) from the abdominal aorta, near the diaphragm, from 15 set before to 2 min after each injection. At the end of the experiment the dog was sacrificed with intravenous KCI. The LLL, the trachea, a section of the left ventricle free wall, and the mid portion of the esophagus were removed and the radioactivity measured in a gamma counter. We calculated the
262
P. G. AGOSTONI TABLE Experimental
Dogs
Sequence
I 2 3 4
30 30 30 30
5 6 7
30 (Ml 30 Of) 30 Of)
(M)
= microspheres
(M) CM) CM) (Ml
of temperatures
er al. 1 protocol
used
33 W 36 36 W 39 Of)
39 33 33 36
CM) CM) W (Ml
33 Of) 33 CM) 33 Of)
39 CM) 36 (Ml 36 (Ml
36 39 39 33
W Of) Of) CM)
36 CM) 39 (Ml 39 (Ml
42 (Ml 42 42 42 (MI 42 w 42 CM)
injection.
systemic flow to each tissue on the basis of the reference flow, the radioactivity of the target tissue and the reference samples (Baile et al., 1982; Heymann et al., 1977). Experimentalprotocol. During the entire experiment, the independent variable was the temperature in the pulmonary artery (Tpa-LLL). After the preparation was completed we collected Qbr,-, with Tpa-LLL at 30 “C until it was stable and for a further 30 min after stabilization. Then we measured Qbr,+ at the other levels of Tpa-LLL for 30 min each. Beginning with 30 “C was necessary because the preparatory surgery was done with the LLL exposed to room temperature. We performed the 42 “C study last in all the experiments because of concern that this high temperature might have slowly reversible effects. In between, the 33, 36 and 39 “C measurements were performed in a random order. Microsphere injections were performed at the end of each temperature period in 3 dogs and at the end of the 4 periods in 4 dogs (table 1).
The data are presented as mean f SE except as noted and analyzed by Student’s paired t-test with the Bonferroni correction such that P < 0.00125 was considered significant (Miller, 1966). The correlation between Qbr (microspheres) and Qbr,-, (overflow) has been analyzed by linear regression. Statistical analysis.
Results Figure 1 illustrates the protocol with the results from one dog. The initial measurements were with Tpa-LLL of 30 “C and the final 42 “C. Qbr,-, is expressed in ml.min-’ .(lOOg dry lung wt)-‘.(lOOTorr‘. The time interval between each temperature step was less than 5 min. As seen in fig. 1, Qbr,-, quickly reached the new level and stabilized. LLL weight was nearly constant throughout the experiment so that only trivial corrections of the overflow volume were necessary to calculate Qbr,.,. Qbr+,, of all dogs are reported for each dog in fig. 2A and as mean f SE in fig. 2B. Qbr
BRONCHIAL
TPWLLL
BLOOD
FLOW
AND
LLL
,
4236-
,o
263
TEMPERATURE
.
t
363330-
5
60-
PP 70E, 605
50-
E s E
403o20-
i
ib
1'5 20 is
30
lb
15 20 25 io
ib
1'5 2b is
jo
ib
1'5 io
is
0
I 5
I I 10 15
I I I 20 25 30
Fig, I. Anastomotic bronchial blood flow (obr,.,) (ml ‘min’ (100 g dry lobe wt‘) corrected to 100 Torr mean systemic pressure, in one dog (dog 2). Temperatures in the left lower lobe pulmonary artery (Tpa-LLL) were, successively, 30, 36, 33, 39, 42 “C. Each Tpa-LLL was maintained for 30 min. Gbr,., collections took 5 min each. Microspheres (M) were injected at the end of all except the 36 “C period.
(microspheres) and Gbr,-, (overflow f weight changes) were measured simultaneously 31 times. They correlated closely: obr = 2.745 t 0.942. Qbrs+, R = 0.904, P < 0.001, SE of intercept = 2.995. Mean systemic blood pressure (Psa), Ppa-LLL, mean LLL airway pressure (Paw-LLL) and the Trpa measured with the Swan-Ganz remained constant throughout the experiment (table 2). When the Tpa-LLL was 39 “C or greater, the blood lost heat through the LLL while, at lower Tpa-LLL, the blood gamed heat in the LLL (table 2). No systematic changes in ot, otr, Cjh or oe occurred in parallel with the change in obr and obr,+ (fig. 3). TABLE
Psa (mm Hg) Ppa-LLL (cm H,O) Paw-LLL (cm H,O) Trpa (“C) DTpv-pa (“C)
2
30°C (n = 7)
33°C (n = 7)
36°C (n = 7)
39°C (n = 7)
42°C (n = 6)
133 20.6 6.4 38.1 0.6
139 20.9 6.4 38.2 0.4
140 19.9 6.5 38.1 0.1
139 20.7 6.4 38.1 -0.3
132 22.3 6.5 38.6 - 0.4
f 20 + 4.4 f 0.2 f I.1 f. 0.1
i: 18 + 4.4 f 0.2 f 1.0 + 0.1
+ 13 + 5.1 f. 0.2 + 0.8 * 0.1
+ 13 + 5.5 + 0.4 f 0.9 + 0.2
+ 19 f 6.7 t 0.3 + 0.8 + 0.2
Data are mean + SD, Psa = mean systemic blood pressure, Ppa-LLL = left lower lobe pulmonary artery pressure, Paw-LLL = mean airway pressure in the left lower lobe, Trpa = temperature in the right pulmonary artery, DTpv-pa = temperature difference between left lower lobe pulmonary vein and artery.
P. G. AGOSTONI
264
et 01. A
3 6o 2 ; D 8 f
.AE E
60-
40-
20-
I
ov/ 30’
36
I
39
I
42
‘c
TDa-LLL
B
6brS-p,oo
1
5 6o P ;
60-
0 8 $ .LE E
40-
* I
;
l
20-
l
F t
I
oh/
’
30
I
I
I
I
33
36
39
42
OC
Tpa-LLL Fig. 2. (A) Anastomotic bronchial blood flow (obr,-,) (ml.min-’ .(I00 g dry lobe wt- ‘) corrected to 100 Torr mean systemic pressure, in each dog (collected for last 5 min of each experimental period). (B) Same plotted as mean f SE: * indicates P -z 0.00125 from Tpa-LLL = 36 “C; + n = 6.
Discussion
Our study shows that the temperature of the left lower lobe pulmonary artery blood influences obr and obrsmp, that the greatest obr and obr,-, occur with LLL temperature near normal (36 ‘C) and that these flow changes are local and do not affect the bronchial blood flow distributed elsewhere, specifically to the trachea and the esophagus. We also confirmed that obr and obr,-, are equal over the range of flow we studied. To account for possible time-dependent variables in the preparation we randomized the temperature sequence between 33 and 39 “C in respect to time. However, our
BRONCHIAL
BLOOD
FLOW
AND
LLL
265
TEMPERATURE
I I h-71
i
(n-71
(n-61
I
(n-71
h-61
(n- 41
(n-6)
I (n-41
(n-6)
(n-41
I (n=71
i
i
h-7)
I
I
I
I
I
30
33
36
39
42
‘c
T Da-LLL Fig. 3. Cardiac output (ot) (ml/min) and flow to the following tissues in units of ml. min- ’ . (100 g dry lobe wt‘: left ventricle free wall (oh), trachea (@r), and the mid portion of the esophagus (be), at pulmonary artery temperatures (Tpa-LLL) of 30, 33, 36, 39, and 42 “C.
preparation was stable throughout the entire experiment as shown by the constancy of Trpa, dog and LLL flows and pressures (fig. 3 and table 2). Lobar obr determined with microspheres (Malik and Tracy, 1980; Baile et al., 1982, 1984, 1985) requires a correction for recirculating microspheres. In our preparation, however, this correction is not necessary because the LLL circulation was isolated and, therefore, no microspheres could reach the lung parenchyma via the pulmonary
266
P. G. AGOSTONI
et al.
circulation. In a preliminary pilot series of dogs we verified that, in our preparation, no microspheres were detectable in reference samples withdrawn from the LLL pulmonary vein from 15 set before to 2 min after the microsphere injection. Our data confirmed those of Baile et al. (1984) who showed in a preparation without pulmonary flow that obr to the LLL was approximately equal to obr,-,. Inside the LLL, non-anastomotic bronchial flow draining into the right heart is negligible in the dog, in accord with the absence of an intraparenchymal true bronchial vein (Evans and Christenen, 1979). Changing the LLL blood temperature from 30 to 42 ‘C did not influence the Ppa-LLL at constant flow. Similar results have been reported by Pennington and colleagues with an even greater reduction of Tpa-LLL (Pennington et al., 1971). On the other hand, Benumof and Wahrenbrock (1977), found that the pulmonary vascular resistance of the LLL increased with whole body hypothermia: they changed the temperature of the entire dog while we only changed temperature in the LLL. At 30 “C they measured a LLL flow (OLLL) similar to our pump flow (260 ml/min). Qt and OLLL in the Benumof and Wahrenbrock study fell as the temperature of the dog fell. The lack of a cold-induced pulmonary vasoconstriction observed in our experiment suggested that, between 30 and 42 “C, temperature does not directly affect pulmonary arteriolar tone. This applies also to lower temperatures according to Pennington et al. (197 1). These observations are also supported by the data of Stem and Braun (1970) who showed that pulmonary vasoconstriction with total body hypothermia is mainly due to an alpha-receptor mediated reflex. In our study, pulmonary artery pressure and blood flow were stable. Therefore, changes in Gbr,-, driving pressure due to differences in pulmonary vascular downstream pressure were not involved in the flow changes we saw. Neither were changes in airway pressure, since this was also constant. The mid portion of the esophagus, the trachea, the bronchi and the bronchopulmonary anastomosis are supplied in common (Evans and Christenen, 1979) from the bronchoesophageal arteries in dogs. Since ot, Gtr, Gh and Qe remained constant through the experiment, the obr and Qbr,-, changes observed were due to local changes in vascular resistance. In particular, the stability of otr and oe shows that bronchial blood flow to the LLL and trachea and esophagus are functionally independent, at least under these circumstances. In our study, we maintained inspired LLL air temperature at 30 ‘C and kept humidity high. Thus we were far from the low humidity, cold dry air conditions previously shown to alter tracheal blood flow (Baile et al., 1985), possibly as a secondary consequence of drying of the mucosa. The pattern of obr and obr,-, responses to Tpa-LLL changes is surprising if one expects to find a positive correlation between temperature and blood flow as is seen in that best known example of temperature-sensitive vascular smooth muscle, the arterioles of human skin (Barcroft and Edholm, 1943). However, Hales et al. (1985) described a temperature dependence of blood flow in sheep skin similar to what we observed. The changes in skin blood flow they described were mainly due to changes in the arteriovenous anastomotic flow. No teleological explanation of our results and those of Hales et al. (1985) comes readily to mind. How could this temperature-dependence of flow
BRONCHIAL
BLOOD
FLOW
AND
LLL
TEMPERATURE
267
exert a protective role? We cannot say. Nonetheless, in the dog, a panting animal, bronchial vascular resistance is strongly influenced by temperature with a minimum near 36 “C.
Acknowledgements.
Dr. John ment.
Butler
The authors for the laboratory
are grateful to J. M. Mendenhall for invaluable technical support and to space, the careful review of the manuscript and the constant encourage-
References Agostoni, P. G., M. E. Deffebach, W. Kirk, J. M. Mendenhall, R. K. Albert and J. Butler (1987). Temperature alters the bronchial blood flow response to pulmonary artery obstruction. J. Appl. Physiol. (in press). Aitken, M. J. and J. J. Marini (1985). Effect of heat delivery and extraction on airway conductance in normal an asthmatic subjects. Am. Rev. Respir. Lb. 131: 357-361. Baier, H., W.M. Long and A. Wanner (1985). Bronchial circulation in asthma. Respiration 48: 199-295. Baile, E. M., B. J. M. Nelems, M. Schultzer and P. Pare (1982). Measurements of regional bronchial arterial blood flow and bronchovascular resistance in dogs. J. Appl. Physiol. 53: 1044-1049. Baile, E. M.. R. Albert, W. Kirk, S. Lakshminarayan, B.J.R. Wiggs and P.D. Pare (1984). Positive end-expiratory pressure decreases bronchial blood flow in the dog. J. Appl. Physiol. 56: 1289-1293. Baile, E. M., R. W. Dahlby, B. R. Wiggs and P. Pare (1985). Role of tracheal and bronchial circulation in respiratory heat exchange. J. Appl. Physiol. 58: 217-222. Barcroft, H. and O.G. Edholm (1943). The effect of temperature on blood flow and deep temperature in the human forearm. J. Physiol. (London) 102: 5-20. Benumof, J.L. and E.A. Wahrenbrock (1977). Dependency of hypoxic pulmonary vasoconstriction on temperature. J. Appl. Physiol. 42: 56-58. Evans, H. and G. C. Christenen (1979). Miller’s Anatomy of the Dog. Philadelphia, London, Toronto, W.B. Saunders Company. Hales, J.R. S., C. Jessen, A.A. Fawcett and R.B. King (1985). Skin AVA and capillary dilatation and constriction induced by local skin heating. Pfiigers Arch. 404: 203-207. Heymann, M. A., B. D. Payne, J. I. E. Hoffman and A. B. Rudolph (1977). Blood flow measurements with radionuclide-labeled particles. Prog. Curdiovasc. Dir. 20: 55-79. Jindal, S. K., S. Lakshminarayan, W. Kirk and J. Butler (1984). Acute increase in anastomotic bronchial blood flow after pulmonary artery obstruction. J. Appl. Physiol. 57: 424-428. Malik, A.B. and S.E. Tracy (1980). Bronchovascular adjustments after pulmonary embolism. J. Appl. Physiol. 49: 476-48 I. McFadden, E. R., D. M. Denison, J. F. Wailer, B. Assouli and A. Peacock (1982). Direct recordings of the temperatures in the tracheobronchial tree in normal man. J. Clin. Invesf. 69: 700-705. McFadden, E. R. (1983). Respiratory heat and water exchange: physiological and clinical implications. J. Appl. Physiol. 54: 331-336. Miller, R.G. (1966). Simultaneous Statistical Inference. New York, McGraw-Hill. Pennington, D. G., A. L. Hyman and W. C. Wooverton (1971). Pulmonary vascular responses to selective lung cooling in intact dogs. Proc. Sot. Exp. Biol. Med. 137: 1375-1380. Stern, D. and K. Braun (1970). Pulmonary arterial and venous response to cooling: role of alpha-adrenergic receptors. Am. J. Physiol. 219: 982-985.
Physiology
(1987)
259
68, 259-267
RSP 01288
Temperature dependence of intraparenchymal bronchial blood flow PiergiuseppeAgostoni, Mark E. Deffebach, Wayne Kirk and George L. Brengelmann Department
of Medicine,
Division of Respiratory Diseases and University of Washington. Seattle, (Accepted
for publication
Department WA, U.S.A.
11 February
of Physiology
and
Biophysics,
1987)
Abstract. Previous studies suggested that bronchial vascular resistance, like that of the skin, changes with the temperature of the surrounding tissue. To investigate this phenomenon, we recorded anastomotic (systemic to pulmonary) (obr,.,) and total (obr) bronchial blood flow over a temperature range centered on normal. In 7 open-chested dogs the in sifu left lower lobe (LLL) was separately ventilated (30 “C, 5% CO, in humidified air) and was suspended in a fabric net from a strain gauge for continuous recording of weight. The pulmonary circulation of the LLL was pump-perfused at 255 f 69 ml/min in a closed circuit with temperature set at 30, 33.36, 39 and 42 “C. obr,., was measured as overflow from the LLL vascular circuit corrected for LLL weight changes. obr, tracheal, mid-esophageal and coronary flow were measured with 15 /J radiolabelled microspheres injected in the left atrium. The animal’s core temperature and that of the humidified air around the LLL were held constant. obr and obr,+ were equal and reached a peak at 36 “C with lower levels of flow at higher and lower temperatures. Esophageal, tracheal and coronary flow and cardiac output did not change nor did pressures in the systemic and LLL pulmonary artery and in the LLL airways. An intralobar change in temperature above or below 36 “C decreases only the lobar bronchial blood flow and does not influence blood flow to other nearby tissues including those vascularized by the bronchial circulation.
Dog;
Esophageal
blood
flow;
Left
lower
lobe;
Microspheres;
Tracheal
blood
flow
Blood flow in the bronchial vasculature appears to be strongly influenced by temperature. In a recent study of the effects of pulmonary artery occlusion (Agostoni et af., 1987) we showed that the discrepancy between measurements of intrapulmonary bronchial blood flow, both anastomotic (obr,-,) and total (Gbr), obtained in an open (Baile et al., 1982, 1985; Jindal et al., 1984) vs a closed chest preparation (Malik and Tracy, 1980) could be explained on the basis of the differing thermal conditions. The physiological significance of this temperature-dependence could stem from the fact that temperatures of mucosa of the trachea and bronchi are influenced by the external Correspondence Italy. 0034-5687/87/$03.50
address:
0
Dr.
P.G.
Agostoni,
1987 Elsevier
Science
Istituto
di Cardiologia,
Publishers
Universita
B.V. (Biomedical
di Milano,
Division)
20138
Milan,
260
P. G. AGOSTONI
ef al.
thermal environment, particularly during hyperventilation with cold air (McFadden et al., 1982). Temperature-dependence of bronchial blood flow could be due to direct thermal effects on tone in vascular smooth muscle of the large airways. A thermally mediated decrease in bronchial blood flow was suggested by the observation that the tracheal mucosa of human subjects turned pale during ventilation with cold air (McFadden, 1983). The opposite, an increase in flow with reduced temperature, has been observed (Baile et al., 1985) in the bronchial circulation of open-chested dogs hyperventilated with cold dry air ( - 20 “C). The interpretation of findings obtained with cold dry air ventilation is complicated by the likelihood that dehydration of the bronchial mucosa by the low humidity of the air may influence the airways’ muscular tone and, indirectly, the bronchial blood flow (A&ken and Marini, 1985; Baier et&, 1985). Temperature-dependent vascular tone could also be a characteristic of bronchial vessels of tissues other than the major airways. Within the pulmonary parenchyma, the influence of pulmonary arterial temperature would be dominant, but bronchial vessel temperature would in part depend upon direct thermal affects of ventilation and in part upon the temperature of arterial blood that arrives via pathways adjacent to proximal larger airways where the air may be at temperatures different from body temperature (McFadden et al., 1982). We designed our experiment to study the temperature dependence of intrapulmonary bronchial blood flow over a range of temperatures established within the parenchyma of a pump-perfused in situ left lower lobe (LLL) in an open-chested dog with the remainder of the preparation held normothermic. With this design, low lobar temperatures could be obtained without the airway drying that would have been inevitable with ventilatory heat exchange. We also used radioactive microspheres to measure flow to the trachea and the middle portion of the esophagus. In the dog, both of these are vascularized, at least partially, by the bronchial circulation (bronchoesophageal arteries; Evans and Christenen, 1979). The intent was to determine whether manipulation of temperature of the isolated lobe would affect flow in these other structures which remained at core temperature.
Methods
Our experimental design called for measurement of systemic blood flow to the trachea (otr), the esophagus toe), the heart (oh), the entire left lower lobe (obr), and the anastomotic, i.e. systemic to pulmonary bronchial blood flow (obr,-,) to the left lower lobe (LLL) with body temperature constant and with the LLL pulmonary artery blood at one of 5 selected temperatures. otr, Qe, oh and Qbr were determined with radiolabelled microspheres using a standard reference flow technique (Heymann et al., 1977) while obr,-, was measured as overflow from the isolated perfused LLL corrected for weight changes (Jindal et al., 1984). Seven mongrel dogs (24.9 + 1.7 kg, mean & SD) were anesthetized with thiopenthal
BRONCHIAL
BLOOD
FLOW
AND
LLL
TEMPERATURE
261
(30 mg/kg) followed by alpha-choralose (30 mg/kg) mixed with urethane (100 mg/kg). The LLL was isolated in situ using a technique previously described (Jindal et al., 1984). The left hemithorax was widely opened and the left upper lobe and the lingula were excised. The right lung and the LLL were ventilated separately with humidified gases (right lung 100% 0, and LLL 5% CO, in air) at a temperature of 30 “C with the end-expiratory pressure held at 5 cm H,O. LLL tidal volume was 150 cc at 12 respirations/min. Without displacing the hilar structures, the LLL was placed in a fabric net hung from a strain gauge. The LLL pulmonary circulation was isolated by inserting glass cannulas in the pulmonary artery and, via the left atrium, in the pulmonary vein. When, rarely, there was more than one vein, we cannulated the larger. We tied off the smaller and relied on the venous anastomoses for its drainage. The extracorporeal LLL vascular circuit included a reservoir with an overflow tube open to the atmosphere at the level of the lower-most portion of the LLL. This reservoir fed a roller pump (0 = 255 f 69 ml/min, mean k SD) and a coil of tubing immersed in a water bath in which the reservoir was also immersed. The LLL pump flow was slowly increased up to 350 ml/mm and/or to a LLL pulmonary artery pressure (Ppa-LLL) of 25 cm H,O. The water bath temperature was set to control the perfusate (autologous blood) temperature. To protect the dog from cooling, the thorax and abdomen were enclosed in a plexiglass box in which the air was fully humidified and maintained between 38 and 39 “C. The back of the dog was thermally isolated from the operating table with an insulating pad. We continuously recorded the LLL weight (strain gauge), the right pulmonary temperature (Trpa, Swan-Ganz), the LLL main pulmonary artery and vein temperature (thermocouples) and pressures (measured at end-expiration with zero reference at the lowermost portion of the LLL). Every 15 min, cardiac output (ot) was measured in triplicate by thermodilution. Blood gases, arterial pH and the hemodynamic status were maintained in the normal range with intravenous fluids and bicarbonate, and by adjusting the right lung ventilation. obr,-, to the LLL was measured as the volume of blood collected from the overflow tube over 5 min. To account for accumulation or loss of intra- or extra-vascular fluids in the lobe, the overflow volume was corrected for the LLL weight changes (Jindal et al., 1984). obr,-, was collected continuously for the entire experiment. obr, Qtr, oh and oe were measured with a reference flow technique (Heymann et al., 1977) by injecting approximately 2 x lo6 radiolabelled microspheres (46Sc, 95Nb, lo3Ru, “‘Sn and 14’Ce) in random order into the left atrium. The diameter of the spheres ranged from 16.5 + 0.1 pm to 15.9 + 0.5 pm. Prior to the injection the spheres were suspended by mechanical mixing and an ultrasonic bath. The reference flow was withdrawn with a constant flow pump (4.6 ml/min) from the abdominal aorta, near the diaphragm, from 15 set before to 2 min after each injection. At the end of the experiment the dog was sacrificed with intravenous KCI. The LLL, the trachea, a section of the left ventricle free wall, and the mid portion of the esophagus were removed and the radioactivity measured in a gamma counter. We calculated the
262
P. G. AGOSTONI TABLE Experimental
Dogs
Sequence
I 2 3 4
30 30 30 30
5 6 7
30 (Ml 30 Of) 30 Of)
(M)
= microspheres
(M) CM) CM) (Ml
of temperatures
er al. 1 protocol
used
33 W 36 36 W 39 Of)
39 33 33 36
CM) CM) W (Ml
33 Of) 33 CM) 33 Of)
39 CM) 36 (Ml 36 (Ml
36 39 39 33
W Of) Of) CM)
36 CM) 39 (Ml 39 (Ml
42 (Ml 42 42 42 (MI 42 w 42 CM)
injection.
systemic flow to each tissue on the basis of the reference flow, the radioactivity of the target tissue and the reference samples (Baile et al., 1982; Heymann et al., 1977). Experimentalprotocol. During the entire experiment, the independent variable was the temperature in the pulmonary artery (Tpa-LLL). After the preparation was completed we collected Qbr,-, with Tpa-LLL at 30 “C until it was stable and for a further 30 min after stabilization. Then we measured Qbr,+ at the other levels of Tpa-LLL for 30 min each. Beginning with 30 “C was necessary because the preparatory surgery was done with the LLL exposed to room temperature. We performed the 42 “C study last in all the experiments because of concern that this high temperature might have slowly reversible effects. In between, the 33, 36 and 39 “C measurements were performed in a random order. Microsphere injections were performed at the end of each temperature period in 3 dogs and at the end of the 4 periods in 4 dogs (table 1).
The data are presented as mean f SE except as noted and analyzed by Student’s paired t-test with the Bonferroni correction such that P < 0.00125 was considered significant (Miller, 1966). The correlation between Qbr (microspheres) and Qbr,-, (overflow) has been analyzed by linear regression. Statistical analysis.
Results Figure 1 illustrates the protocol with the results from one dog. The initial measurements were with Tpa-LLL of 30 “C and the final 42 “C. Qbr,-, is expressed in ml.min-’ .(lOOg dry lung wt)-‘.(lOOTorr‘. The time interval between each temperature step was less than 5 min. As seen in fig. 1, Qbr,-, quickly reached the new level and stabilized. LLL weight was nearly constant throughout the experiment so that only trivial corrections of the overflow volume were necessary to calculate Qbr,.,. Qbr+,, of all dogs are reported for each dog in fig. 2A and as mean f SE in fig. 2B. Qbr
BRONCHIAL
TPWLLL
BLOOD
FLOW
AND
LLL
,
4236-
,o
263
TEMPERATURE
.
t
363330-
5
60-
PP 70E, 605
50-
E s E
403o20-
i
ib
1'5 20 is
30
lb
15 20 25 io
ib
1'5 2b is
jo
ib
1'5 io
is
0
I 5
I I 10 15
I I I 20 25 30
Fig, I. Anastomotic bronchial blood flow (obr,.,) (ml ‘min’ (100 g dry lobe wt‘) corrected to 100 Torr mean systemic pressure, in one dog (dog 2). Temperatures in the left lower lobe pulmonary artery (Tpa-LLL) were, successively, 30, 36, 33, 39, 42 “C. Each Tpa-LLL was maintained for 30 min. Gbr,., collections took 5 min each. Microspheres (M) were injected at the end of all except the 36 “C period.
(microspheres) and Gbr,-, (overflow f weight changes) were measured simultaneously 31 times. They correlated closely: obr = 2.745 t 0.942. Qbrs+, R = 0.904, P < 0.001, SE of intercept = 2.995. Mean systemic blood pressure (Psa), Ppa-LLL, mean LLL airway pressure (Paw-LLL) and the Trpa measured with the Swan-Ganz remained constant throughout the experiment (table 2). When the Tpa-LLL was 39 “C or greater, the blood lost heat through the LLL while, at lower Tpa-LLL, the blood gamed heat in the LLL (table 2). No systematic changes in ot, otr, Cjh or oe occurred in parallel with the change in obr and obr,+ (fig. 3). TABLE
Psa (mm Hg) Ppa-LLL (cm H,O) Paw-LLL (cm H,O) Trpa (“C) DTpv-pa (“C)
2
30°C (n = 7)
33°C (n = 7)
36°C (n = 7)
39°C (n = 7)
42°C (n = 6)
133 20.6 6.4 38.1 0.6
139 20.9 6.4 38.2 0.4
140 19.9 6.5 38.1 0.1
139 20.7 6.4 38.1 -0.3
132 22.3 6.5 38.6 - 0.4
f 20 + 4.4 f 0.2 f I.1 f. 0.1
i: 18 + 4.4 f 0.2 f 1.0 + 0.1
+ 13 + 5.1 f. 0.2 + 0.8 * 0.1
+ 13 + 5.5 + 0.4 f 0.9 + 0.2
+ 19 f 6.7 t 0.3 + 0.8 + 0.2
Data are mean + SD, Psa = mean systemic blood pressure, Ppa-LLL = left lower lobe pulmonary artery pressure, Paw-LLL = mean airway pressure in the left lower lobe, Trpa = temperature in the right pulmonary artery, DTpv-pa = temperature difference between left lower lobe pulmonary vein and artery.
P. G. AGOSTONI
264
et 01. A
3 6o 2 ; D 8 f
.AE E
60-
40-
20-
I
ov/ 30’
36
I
39
I
42
‘c
TDa-LLL
B
6brS-p,oo
1
5 6o P ;
60-
0 8 $ .LE E
40-
* I
;
l
20-
l
F t
I
oh/
’
30
I
I
I
I
33
36
39
42
OC
Tpa-LLL Fig. 2. (A) Anastomotic bronchial blood flow (obr,-,) (ml.min-’ .(I00 g dry lobe wt- ‘) corrected to 100 Torr mean systemic pressure, in each dog (collected for last 5 min of each experimental period). (B) Same plotted as mean f SE: * indicates P -z 0.00125 from Tpa-LLL = 36 “C; + n = 6.
Discussion
Our study shows that the temperature of the left lower lobe pulmonary artery blood influences obr and obrsmp, that the greatest obr and obr,-, occur with LLL temperature near normal (36 ‘C) and that these flow changes are local and do not affect the bronchial blood flow distributed elsewhere, specifically to the trachea and the esophagus. We also confirmed that obr and obr,-, are equal over the range of flow we studied. To account for possible time-dependent variables in the preparation we randomized the temperature sequence between 33 and 39 “C in respect to time. However, our
BRONCHIAL
BLOOD
FLOW
AND
LLL
265
TEMPERATURE
I I h-71
i
(n-71
(n-61
I
(n-71
h-61
(n- 41
(n-6)
I (n-41
(n-6)
(n-41
I (n=71
i
i
h-7)
I
I
I
I
I
30
33
36
39
42
‘c
T Da-LLL Fig. 3. Cardiac output (ot) (ml/min) and flow to the following tissues in units of ml. min- ’ . (100 g dry lobe wt‘: left ventricle free wall (oh), trachea (@r), and the mid portion of the esophagus (be), at pulmonary artery temperatures (Tpa-LLL) of 30, 33, 36, 39, and 42 “C.
preparation was stable throughout the entire experiment as shown by the constancy of Trpa, dog and LLL flows and pressures (fig. 3 and table 2). Lobar obr determined with microspheres (Malik and Tracy, 1980; Baile et al., 1982, 1984, 1985) requires a correction for recirculating microspheres. In our preparation, however, this correction is not necessary because the LLL circulation was isolated and, therefore, no microspheres could reach the lung parenchyma via the pulmonary
266
P. G. AGOSTONI
et al.
circulation. In a preliminary pilot series of dogs we verified that, in our preparation, no microspheres were detectable in reference samples withdrawn from the LLL pulmonary vein from 15 set before to 2 min after the microsphere injection. Our data confirmed those of Baile et al. (1984) who showed in a preparation without pulmonary flow that obr to the LLL was approximately equal to obr,-,. Inside the LLL, non-anastomotic bronchial flow draining into the right heart is negligible in the dog, in accord with the absence of an intraparenchymal true bronchial vein (Evans and Christenen, 1979). Changing the LLL blood temperature from 30 to 42 ‘C did not influence the Ppa-LLL at constant flow. Similar results have been reported by Pennington and colleagues with an even greater reduction of Tpa-LLL (Pennington et al., 1971). On the other hand, Benumof and Wahrenbrock (1977), found that the pulmonary vascular resistance of the LLL increased with whole body hypothermia: they changed the temperature of the entire dog while we only changed temperature in the LLL. At 30 “C they measured a LLL flow (OLLL) similar to our pump flow (260 ml/min). Qt and OLLL in the Benumof and Wahrenbrock study fell as the temperature of the dog fell. The lack of a cold-induced pulmonary vasoconstriction observed in our experiment suggested that, between 30 and 42 “C, temperature does not directly affect pulmonary arteriolar tone. This applies also to lower temperatures according to Pennington et al. (197 1). These observations are also supported by the data of Stem and Braun (1970) who showed that pulmonary vasoconstriction with total body hypothermia is mainly due to an alpha-receptor mediated reflex. In our study, pulmonary artery pressure and blood flow were stable. Therefore, changes in Gbr,-, driving pressure due to differences in pulmonary vascular downstream pressure were not involved in the flow changes we saw. Neither were changes in airway pressure, since this was also constant. The mid portion of the esophagus, the trachea, the bronchi and the bronchopulmonary anastomosis are supplied in common (Evans and Christenen, 1979) from the bronchoesophageal arteries in dogs. Since ot, Gtr, Gh and Qe remained constant through the experiment, the obr and Qbr,-, changes observed were due to local changes in vascular resistance. In particular, the stability of otr and oe shows that bronchial blood flow to the LLL and trachea and esophagus are functionally independent, at least under these circumstances. In our study, we maintained inspired LLL air temperature at 30 ‘C and kept humidity high. Thus we were far from the low humidity, cold dry air conditions previously shown to alter tracheal blood flow (Baile et al., 1985), possibly as a secondary consequence of drying of the mucosa. The pattern of obr and obr,-, responses to Tpa-LLL changes is surprising if one expects to find a positive correlation between temperature and blood flow as is seen in that best known example of temperature-sensitive vascular smooth muscle, the arterioles of human skin (Barcroft and Edholm, 1943). However, Hales et al. (1985) described a temperature dependence of blood flow in sheep skin similar to what we observed. The changes in skin blood flow they described were mainly due to changes in the arteriovenous anastomotic flow. No teleological explanation of our results and those of Hales et al. (1985) comes readily to mind. How could this temperature-dependence of flow
BRONCHIAL
BLOOD
FLOW
AND
LLL
TEMPERATURE
267
exert a protective role? We cannot say. Nonetheless, in the dog, a panting animal, bronchial vascular resistance is strongly influenced by temperature with a minimum near 36 “C.
Acknowledgements.
Dr. John ment.
Butler
The authors for the laboratory
are grateful to J. M. Mendenhall for invaluable technical support and to space, the careful review of the manuscript and the constant encourage-
References Agostoni, P. G., M. E. Deffebach, W. Kirk, J. M. Mendenhall, R. K. Albert and J. Butler (1987). Temperature alters the bronchial blood flow response to pulmonary artery obstruction. J. Appl. Physiol. (in press). Aitken, M. J. and J. J. Marini (1985). Effect of heat delivery and extraction on airway conductance in normal an asthmatic subjects. Am. Rev. Respir. Lb. 131: 357-361. Baier, H., W.M. Long and A. Wanner (1985). Bronchial circulation in asthma. Respiration 48: 199-295. Baile, E. M., B. J. M. Nelems, M. Schultzer and P. Pare (1982). Measurements of regional bronchial arterial blood flow and bronchovascular resistance in dogs. J. Appl. Physiol. 53: 1044-1049. Baile, E. M.. R. Albert, W. Kirk, S. Lakshminarayan, B.J.R. Wiggs and P.D. Pare (1984). Positive end-expiratory pressure decreases bronchial blood flow in the dog. J. Appl. Physiol. 56: 1289-1293. Baile, E. M., R. W. Dahlby, B. R. Wiggs and P. Pare (1985). Role of tracheal and bronchial circulation in respiratory heat exchange. J. Appl. Physiol. 58: 217-222. Barcroft, H. and O.G. Edholm (1943). The effect of temperature on blood flow and deep temperature in the human forearm. J. Physiol. (London) 102: 5-20. Benumof, J.L. and E.A. Wahrenbrock (1977). Dependency of hypoxic pulmonary vasoconstriction on temperature. J. Appl. Physiol. 42: 56-58. Evans, H. and G. C. Christenen (1979). Miller’s Anatomy of the Dog. Philadelphia, London, Toronto, W.B. Saunders Company. Hales, J.R. S., C. Jessen, A.A. Fawcett and R.B. King (1985). Skin AVA and capillary dilatation and constriction induced by local skin heating. Pfiigers Arch. 404: 203-207. Heymann, M. A., B. D. Payne, J. I. E. Hoffman and A. B. Rudolph (1977). Blood flow measurements with radionuclide-labeled particles. Prog. Curdiovasc. Dir. 20: 55-79. Jindal, S. K., S. Lakshminarayan, W. Kirk and J. Butler (1984). Acute increase in anastomotic bronchial blood flow after pulmonary artery obstruction. J. Appl. Physiol. 57: 424-428. Malik, A.B. and S.E. Tracy (1980). Bronchovascular adjustments after pulmonary embolism. J. Appl. Physiol. 49: 476-48 I. McFadden, E. R., D. M. Denison, J. F. Wailer, B. Assouli and A. Peacock (1982). Direct recordings of the temperatures in the tracheobronchial tree in normal man. J. Clin. Invesf. 69: 700-705. McFadden, E. R. (1983). Respiratory heat and water exchange: physiological and clinical implications. J. Appl. Physiol. 54: 331-336. Miller, R.G. (1966). Simultaneous Statistical Inference. New York, McGraw-Hill. Pennington, D. G., A. L. Hyman and W. C. Wooverton (1971). Pulmonary vascular responses to selective lung cooling in intact dogs. Proc. Sot. Exp. Biol. Med. 137: 1375-1380. Stern, D. and K. Braun (1970). Pulmonary arterial and venous response to cooling: role of alpha-adrenergic receptors. Am. J. Physiol. 219: 982-985.