PULMONARY ARTERIOVENOUS SHUNTS DURING HALOTHANE ANAESTHESIA IN DOGS*

Br.J. Anaesth. (1980) 52, 763

PULMONARY ARTERIOVENOUS SHUNTS DURING HALOTHANE ANAESTHESIA IN DOGS* D. J. PAVLIN, J. FERENS, D. R. ALLEN AND F. W. CHENEY SUMMARY

General anaesthesia is frequently associated with an increase in the alveolar-arterial oxygen tension gradient ( P A Q J - P A Q , ) and physiological shunt pl fraction (Campbell, Nunn and Peckett, 1958; Stark and d Smith, 1960; Theye and Tuohy, 1964; Panday and Nunn, 1968; Marshall, Cohen and Klingenmaier, 1969; Stone et al., 1975). The intrapuknonary component of physiological shunt may result from blood perfusing non-ventilated alveoli or blood passing through direct arteriovenous (A-V) anastomoses. The size, distribution and physiological role of A-V anastomoses is poorly understood, and their contribution to the total physiological shunt fraction during general anaesthesia unknown. Although often suggested as a potential cause of hypoxia during anaesthesia (Marshall and Wyche, 1972; Rehdcr, Sessler and Marsh, 1975), flow through these channels has not been measured during general anaesthesia and compared with the awake control state. It has been suggested that these channels may open in response to increased pulmonary artery pressure (Cheney et al., 1978), beta-adrenergic stimulation or general anaesthesia. Balchum and others (1968) by direct observation during angiography, reported opening of such channels when pulmonary artery pressure was increased. More D. JANET PAVLIN, M.SC, M,D. ; FREDERICK W. CHENEY, M.D. ; Department of Anesthesiology. JOHN FERENS, B.S. ; DAVID R. ALLEN, PH.D.J Division of Nuclear Medicine, Pharma-

ceutical Sciences and Radiology. University of Washington School of Medicine, Seattle, Washington 98195, U.S.A. * Presented at the American Society of Anesthesiologists Annual Meeting, Chicago, Illinois, 1975. 0007-0912/80/080763-06 $01.00

recently, measurements of A-V shunt flow have been made using radioactively labelled microspheres. Using such quantitative methods, increasing pulmonary artery pressure was found to have no effect on A-V shunt flow (Cheney et al., 1978). Nomoto and others (1974) used similar techniques and reported a five-fold increase in anastomotic shunt flow in dogs during adrenaline infusion, which led them to suggest that these channels may open in response to betaadrenergic stimulation. The purpose of this study was to assess the role of A-V anastomoses as a cause of hypoxia during halothane anaesthesia since halothane might open the channels either directly by causing relaxation of vascular smooth muscle or indirectly by a betaagonist effect (Price et al., 1970). Such a mechanism might partially explain the decrease in P a ^ commonly observed during anaesthesia with halothane and other inhalation agents. The method was designed to measure the fraction of cardiac output shunted through pulmonary A-V channels during halothane anaesthesia compared with the awake control state. METHODS

Group 1. Halothane anaesthesia Anaesthesia was induced with halothane in nine adult unpremedicated male mongrel dogs, weighing 8-38 kg. Each animal was placed supine, the trachea was intubated with a cuffed tube, and anaesthesia was maintained with 1% halothane in oxygen. Ventilation was controlled using a Harvard pump to deliver a tidal volume of 15 ml kg" 1 , and the respiratory frequency adjusted to maintain Pa^ between © Macmillan Publishers Ltd 1980

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on July 28, 2015

The fraction of cardiac output flowing through pulmonary arteriovenous anastomoses (£»v/£)t) was measured in dogs during halothane anaesthesia and compared with results obtained in normal awake dogs. Flow through arteriovenous anastomoses was measured using t8mTc-labelled polystyrene microspheres (50 pun diameter). The fraction of cardiac output perfusing pulmonary arteriovenous anastomoses during halothane anaesthesia was 4.1% ( ± S D 1.75) compared with 4.6% ( ± S D 0.73) in air-breathing awake control dogs. In spite of variations in arteriovenous shunt fraction, no significant relationship between the ()»v/£?t a n ( i (?«/& was detected. These results suggest that pulmonary arteriovenous anastomoses do not contribute significantly to the physiological shunt observed during halothane anaesthesia.

BRITISH JOURNAL OF ANAESTHESIA

764

£)a_ () t

increased count rate after IVC injection increased count rate after LV injection

Ajy ^mc

where A^v and AIVC are the activity of B8m Tclabelled spheres injected into the LV and IVC respectively. All count rates were corrected for

background activity and decay of technetium-99m. Group 2. Awake control The technique described above required the animal to remain motionless during the counting period and was therefore inappropriate for measurement of A-V shunt in the awake state. Therefore, a modification was used to measure £) av in five awake animals (group 2). Radioactive microspheres injected i.v. will either pass through A-V shunts in tie lung and lodge in the microcirculation of the remainder of the body (carcass) or will be trapped in the microcirculation of the lung. A catheter was inserted to a peripheral vein in the forelimb of an awake dog and labelled microspheres injected followed immediately by 10 ml of isotonic saline. The animal was sacrificed after 2 min by injection of sodium thiopentone 2 g i.v. The chest was opened immediately and clamps placed across the pulmonary arteries and veins at the hilum. The pulmonary vessels were severed between clamps so that no blood was allowed to escape from either the proximal or distal ends of the cut vessels. The lungs were then removed and placed in a plastic bag. The radioactivity in the lungs and in the remainder of the body was determined separately using a gamma counter. The A-V shunt fraction was determined as follows (method 2): (counts in carcass) (counts in carcass + lungs)

arteriovenous shunt fraction

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on July 28, 2015

3.7 and 5.3 kPa. Catheters were introduced to the femoral artery, inferior vena cava (IVC), pulmonary artery (PA) and the left ventricle (LV) via the carotid artery. A sodium iodide detector was positioned over the head of the animal and a lead shield placed over the shoulders so that placement of a radioactive sample on the chest resulted in no increase in background counts recorded over the head. The details of the radioactive microsphere technique for measuring A-V anastomotic shunt fraction have been reported previously (Cheney et al., 1978). Microspheres (43-53 pun diameter) labelled with technetium-99m were injected into the left ventricle and the count rate over the head of the dog measured for 5 min using a Picker dual probe sodium iodide detector with a single channel analyser and sealer. A second injection of microspheres was then made into IVC and activity determined over a subsequent 5-min period. A-V anastomotic shunt fraction was calculated as described below. Measurements of pulmonary and systemic vascular pressures, cardiac output, arterial and mixed venous blood-gas tensions, haemoglobin concentration and percent saturation were made in the same time period following each injection of microspheres. Physiological shunt fraction (() e /£) t when FiOt= 1.0) was determined by the oxygen technique using the standard shunt equation. Cardiac output was measured by the dye dilution technique using indocyanine green. Blood-gas tensions were measured with a Radiometer blood-gas analyser. Arteriovenous shunt calculation. A-V shunt was determined by plotting activity in accumulated counts per 30 s against time and extrapolating the data to obtain the count rate 30 s after LV and IVC injections (fig. 1). This interval was chosen arbitrarily to allow time for complete circulation of blood and microspheres from the injection site at IVC to the point of lodging in the head. The total amount of activity injected was determined by measurement before injection (range 50-750 y.d). The anastomotic shunt (£)av/£)t) was calculated as follows (method 1):

It was possible that sacrificing the animal might have aflFected the distribution of microspheres between the lung and the remainder of the body. Three experiments were performed to determine the effects of sacrificing an animal by injection of sodium thiopentone 2 g i.v. In three dogs anaesthetized with halothane, an injection of labelled microspheres was made to the IVC and baseline counts over the head recorded for 5 min. Sodium thiopentone 2 g was then injected i.v. No change in the amount of activity counted over the head was detected after 3 min (P<0.05; Student's t test for non-paired data). It was concluded that this method of sacrificing awake control dogs did not dislodge microspheres trapped previously in the lungs. Group 3. Halothane anaesthetized dogs A-V shunt fraction was measured by method 2 in an additional group of four animals anaesthetized with 1% halothane in 99% oxygen. To determine

A-V SHUNTS DURING HALOTHANE ANAESTHESIA

765

Background activity IVC injection microspheres Data point Extrapolated value 30s after injection 0

2

4

6

12

8 10 (Time rrtin)

14

16

18

20

FIG. 1. Typical artenovenous shunt determination in a dog during halothane anaesthesia g " m Tc-labelled microspheres; A->B = background radioactivity, B=left ventricular injection of microspheres, C = inferior vena cava injection of microspheres; O = radioactive counts x 10*/30-s interval; D = extrapolated counts 30 s after microsphere injection.

whether the two different methods of measuring A-V shunt flow produced comparable results. Anaesthesia was induced by inhalation of 4% halothane in oxygen by mask and maintained by spontaneous ventilation with 1% halothane in oxygen. Labelled microspheres (43-53 UJH diameter) were injected i.v. to a forepaw vein, and flushed with 5 ml of saline. After 2 min, the animal was sacrificed by injection of sodium thiopentone 2 g. Thoracotomy was performed rapidly, proximal and distal clamps placed across both hila and the lungs removed by severing between the clamps. Radioactivity in the remaining carcass and in the lungs was determined separately using a gamma counter. The A-V shunt fraction was then calculated as described previously in the normal awake control group (method 2).

10 "Normal awake" controls

8-

8

Halothane with F\Q 1.0

6

X

S

4

RESULTS

Halothane anaesthesia caused no significant change in the fraction of cardiac output perfusing A-V shunts compared with the awake control state (fig. 2). Mean $av/£?t measured during halothane anaesthesia (group 1) was 4.1% (+ SD 1.75) using method 1, compared with 4.6% ( ± S D 0.73) in the awake control dogs (group 2) using method 2. When Pa Ot and physiological shunt were compared (using correlation

2-

FIG. 2. Artenovenous shunt flow expressed as a percentage of cardiac output in normal awake control dogs (group 2, cross-hatched bars), and during 1% halothane anaesthesia (group 1, stippled bars).

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on July 28, 2015

LV injection microspheres

BRITISH JOURNAL OF ANAESTHESIA

766

TABLE I. Physiological and arteriovenous shunt and cardiovascular results during halothane anaesthesia. MAP= mean arterial pressure, CI= cardiac index, P J w = mean pulmonary artery wedge pressure, PXp= mean pulmonary artery pressure, PVR= pulmonary vascular resistance; r= correlation coefficient between (2 a T /& and haemodynandc variables; nj. = no significant correlation (P>0.05).

PVR

CI (litre

xlOO

1.3 2.7 3.0 3.3 4.3 4.5 4.8 6.2 6.9

10.5 10.5 12.0 11.5 11.0 10.5 10.0 11.0 11.5 11.2

4.1 1.8

SD r

Significance

Tnin~'

PAW

PAP

m-»)

(kPa)

(kPa)

14.2 11.6 14.1 12.8

2.86 1.04 1.94 1.42 4.03 2.09 3.60 2.07 1.45

0 1.3 0 1.3 0.4 0.9 0.7 0.5 0.5

1.3 2.5 2.1 1.2 2.0 2.4 2.2 1.8 2.0

69 74 62 72 68 66 66 68 68

2.28 1.02 -0.02

0.6 0.46 -0.02

2.0 0.4

68 3.5 0

n.s.

n.s.

12.
0.9

14.1 1.5

±0.46 n.s.

±0.62 n.s.

analysis) with simultaneously determined () av /£) t in group 1, no significant relationship was detected. Similar analysis of data failed to reveal any significant relationship between anastomotic shunt fraction and mean pulmonary artery pressure, mean pulmonary artery wedge pressure, pulmonary vascular resistance or cardiac index, suggesting that none of these variables has any direct influence on £ a v / $ t (table I). The mean value of () aT /£) t during halothane anaesthesia measured using method 2 in group 3 was 2.3% ( ± S D 0.23). This was not significantly different (Student's t test for unpaired data) from the values obtained in group 1 animals using method 1 (P>0.05) indicating that the two methods of determining the A-V shunt fraction produce comparable results.

DISCUSSION

The validity of the radioactive microsphere technique for measuring regional distribution of blood flow has been demonstrated previously (Rudolph and Heymann, 1967; Wagner et al., 1969). Using similar techniques, Strauss and others (1969) and Nomoto and colleagues (1974) found 2.2 and 2.1% anastomotic shunt respectively in dogs anaesthetized with pentobarbitone. These figures are slightly less than those obtained in this study and this may be explained by small differences in composition or size of microspheres used in different laboratories. Under conditions similar to those in which

+ 0.26 n.s.

PSQ

n.s.

Pa (kPa) 4.5 4.1 4.0 4.3 4.6 4.6 7.4 7.4 4.0 4.2 0.4 -0.45 n.s.

(kPa litre- 1

min) 0.47 1.22 1.10 0.85 0.50 0.71 0.43 0.74 1.49 0.83 0.37 0

n.s.

halothane is administered clinically, there was no significant change in the fraction of cardiac output flowing through A-V anastomoses compared with awake control values. Conclusions are based on the measurement of A-V shunt fraction with a single size microsphere (43-53 (xm diameter). Spheres of this size were chosen to avoid measurement of blood flow through capillaries which are considered to be 5-15 ysa in diameter (Weibel, 1963). Although flow through shunts with a diameter of less than 50 \im will not be measured by this technique, it is likely that microspheres will be squeezed through normal capillaries when smaller microspheres are chosen as indicators. We have shown previously that values of shunt determined using larger microspheres (110 (un) are near zero (Cheney et al., 1978). It seems likely, therefore, that the 50-(ixn spheres provide an adequate assessment of the response of anastomotic channels to anaesthesia and to variation in haemodynamic variables. The relevance of our results to man is unknown. However, Strauss and others (1969) have found similar values of A-V shunt in man and dog. It therefore seems likely that shunts behave similarly in the two species during anaesthesia. Halothane was chosen for these experiments because of its frequent use in clinical practice and because of its proposed beta-agonist activity. Other volatile agents such as fluoroxene, ether or cyclopropane, which are known to be associated with increased concentrations of catecholamines, might give different results if A-V

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on July 28, 2015

Mean

xlOO

MAP (kPa)

A-V SHUNTS DURING HALOTHANE ANAESTHESIA

ACKNOWLEDGEMENTS

The authors wish to thank Rodney Gronka and Ed Chappelle for their technical assistance. This study was supported in part by USPHS Grants GM15991-07 (Anesthesia Research Center) and GM0116012 (Anesthesia Research Training). REFERENCES

Balchum, O. J., Jung, R. C , Turner, A. F., and Jacobson, G. (1968). Pulmonary artery to vein shunts. Proceedings of Ninth Aspen Emphysema Conference, p. 223. Berk, J. L., Hagen, J. H., Beyer, W. H., and Niazmand, R. (1967). The effect of epinephrine on arteriovenous shunts in the pathogenesis of shock. Surg. Gynecol. Obstet., 124, 347. Campbell, E. J. M., Nunn, J. F., and Peckett, B. W. (1958). A comparison of artificial ventilation and spontaneous respiration with particular reference to ventilation-blood flow relationships. Br. J. Anaesth., 30, 166. Cheney, F. W., Pavlin, J., Ferens, J., and Allen, D. (1978). Effect of pulmonary microembolism on arteriovenous shunt flow. J. Thorac. Cardiovasc. Surg., 76, 473. Marshall, B. E., Cohen, P. J., and Klingenmaier, C. H. (1969). Pulmonary venous admixture before, during, and after halothane in man. J. Appl. Physio!., 27, 653. Wyche, M. Q. (1972). Hypoxemia during and after anesthesia. Anesthesiology, 37, 178. Nomoto, S., Berk, J. L., Hagen, J. F., and Koo, R. (1974). Pulmonary anastomic arteriovenous shunting caused by epinephrine. Arch. Surg., 108, 201. Panday, J., and Nunn, J. F. (1968). Failure to demonstrate progressive falls of arterial P o , during anaesthesia. Anaesthesia, 23, 38. Price, H. L., Skorsted, M. P., Pauca, A. L., and Cooperman, L. H. (1970). Evidence for p-receptor activation produced by halothane in normal man. Anesthesiology. 32, 389. Rehder, K., Sessler, A. D., and March, H. M. (1975). General anesthesia and the lung. Am. Rev. Respir. Dis, 112, 541. Rudolph, A. W., and Heymann, M. A. (1967). The circulation of the fetus in utero. Circ. Res., 21, 163. Stark, D. C , and Smith, H. (1960). Pulmonary vascular changes during anaesthesia. Br. J. Anaesth., 32, 460.

Stone, J. G., Khambatta, H. J., Donham, R. T., and Sullivan, S. F. (1975). Pulmonary shunting during anaesthesia in man. Can. Anaesth. Soc. J., 22, 647. Strauss, H. W., Hurley, P. J., Rhodes, B. A., and Wanger, H. N. (1969). Quantification of right to left transpulmonary shunts in man. J. Lab. Clin. Med., 74, 598. Theye, R. A., and Tuohy, G. F. (1964). Oxygen uptake during light halothane anesthesia in man. Anesthesiology, 25, 627. Wagner, H. N., Rhodes, B. A., Sasaki, V., and Ryan, J. P. (1969). Studies of the circulation with radioactive microspheres. Invest. RadioL, 4, 374. Weibel, E. R. (1963). Morphometry of the Human Lung, p. 82. New York: Academic Press.

DERIVATIONS ARTERIOVEINEUSES PULMONAIRES PENDANT UNE ANESTHESIE A L'HALOTHANE SUR DBS CHIENS RESUME

On a mesurd sur des chiens anesth6si6s a l'halothane la fraction de dibit cardiaque passant a travers des anastomoses artirioveineuses pulmonaires (£) sv /6t) e t o n a compari les r&ultats obtenus avec ceux de chiens normaux eveilles. On a mesure le ddbit passant a travers les anastomoses artirioveineuses a l'aide de microspheres en polystyrene (50 ysn de diametre) marquees " m T c . La fraction de dibit cardiaque passant au travers des anastomoses artirioveineuses pulmonaires pendant l'anesthisie a l'halothane a etc de 4,1% (±ecart type 1,75) par rapport a 4,6% ( ± 6cart type 0,73) chez les chiens timoins eVeillis respirant de l'air. En dipit des variations dans la fraction de derivation artirioveineuse, on n'a diceli aucune relation notoire entre $av/£>t et Q,l@f Ces r&ultats laissent penser que les anastomoses artirioveineuses pulmonaires ne contribuent pas d'une maniere significative a la derivation physiologique observie pendant l'anesthisie a Phalothane.

ARTERIOVENOSE LUNGEN-SHUNTS BEI HALOTHAN-NARKOSE IN HUNDEN ZUSAMMENFASSUNG

Der Bruchteil des Herzminutenvolumens, der durch arteriovenose Lungen-Anastomosen fliesst ((?aT/6t)> wurde bei Hunden wahrend einer Halothannarkose gemessen und mit Resultaten bei normal wachen Hunden vergh'chen. Der Durchfluss durch arteriovenose Anastomosen wurde mittels "mTc-Polystyren-Zentrosomen (50 (jun Durchmesser) gemessen. Der Teil des Herzminutenvolumens in den artcriovenosen Lungen-Anastomosen wahrend der Halothannarkose betrug 4,1% ( ± S D 1,75), verglichen mit 4,6% ( ± S D 0,73) bei luftatmenden wachen Hunden. Trotz der Anderungen im arteriovenosen Shunt-Anteil wurde keine signifikante Beziehung zwischen ^ S T /^t und 6»/6t entdeckt. Die Resultate zeigen, dass arteriovenose Lungen-Anastomosen nicht wesentlich zum physiologiscnen Shunt beitragen, der wShrend der Halothannarkose beobachtet wurde.

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on July 28, 2015

communications are influenced by catecholamines as proposed by Berk and his colleagues (1967). In this and in a previous study we have shown that there is no significant relationship between pulmonary artery pressure and arteriovenous shunt flow which suggests that A-V communications do not open in response to increased pulmonary artery pressure. Despite variation in £)av/£)f there was no correlation with Q^Qtp which suggests that A-V anastomotic shunt flow is not a significant factor in QjQt during halothane anaesthesia. It is likely that physiological shunt observed during halothane anaesthesia is a result of shunting of blood through non-ventilated, collapsed areas of lung.

767

768

BRITISH JOURNAL OF ANAESTHESIA

DERIVACIONES ARTERIOVENOSAS DEL PULMON DE PERROS DURANTE ANESTESIA CON HALOTANO

filtro la anastomosis arteriovenosa de pulmdn durante la anestesia con halotano rue del 4,1% (±desviaci6n tipica de 1,75) en comparaci6n con un valor de 4,6% (±desviaci6n tipica de 0,73) en perros despiertos de control que respiSUMARIO raban aire. A pesar de las variaciones en la fracci6n correSe midi6 en perros la fraccidn del flujo cardiaco que fluye a spondiente a la derivaci6n arteriovenosa, no se detect6 traves de la anastomosis arteriovenosa de pulm6n ((2av/(Jt), relaci6n significativa alguna entre el &T/£?t y el & / & durante la anestesia con halotano y se compard con los Estos resultados sugieren que la anastomosis arteriovenosa resultados obtenidos en perros despiertos y nonnales. El de pulm6n no contribuye de forma significativa a la deriflujo a traves de la anastomosis arteriovenosa se midi6 vation fisiologica observada durante la anestesia con usando microesferas de poliestireno irradiadas (MmTc) halotano. (de 50 (im de diametro). La fracci6n del flujo cardiaco que

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on July 28, 2015