Hemodynamic and regional blood flow distribution responses to dextran, hydralazine, isoproterenol and amrinone during experimental cardiac tamponade

J AM CaLL CARDIOL

1461

1983:1(6) 1461-70

Hemodynamic and Regional Blood Flow Distribution Responses to Dextran, Hydralazine, Isoproterenol and Amrinone During Experimental Cardiac Tamponade RONALD W. MILLARD, PhD, NOBLE O. FOWLER, MD, FACC, MARJORIE GABEL Cincinnati, Ohio

Four different interventions were examined in dogs with cardiac tamponade. Infusion of 216 to 288 ml saline solution into the pericardium reduced cardiac output from 3.5 ± 0.3 to 1.7 ± 0.2 liters/min as systemic vascular resistance increased from 4,110 ± 281 to 6,370 ± 424 dynes-s-cm", Left ventricular epicardial and endocardial blood flows were 178 ± 13 and 220 ± 12 ml/ min per 100 g, respectively, and decreased to 72 ± 14 and 78 ± 11 ml/min per 100 g with tamponade. Reductions of 25 to 65 % occurred in visceral and brain blood flows and in a composite brain sample. Cardiac output during tamponade was significantly increased by isoproterenol, 0.5 ILg/kg per min intravenously; hydralazine, 40 mg intravenously; dextran infusion or combined hydralazine and dextran, but not by amrinone. Total systemic vascular resistance was reduced by all interventions. Left ventricular epicardial flow was in-

Acute cardiac tamponade accompanying traumatic thoracic injury most commonly is treated by needle aspiration or surgical drainage. Chronic, progressive cardiac tamponade associated with pericarditis or neoplastic diseases is most frequently relieved by transthoracic needle aspiration. However, temporary improvement of cardiac output may be of advantage in cases where tamponade cannot be relieved without surgical drainage of the pericardial sac; therefore,

From the Departments of Pharmacology and Cell Biophysics. and Medicine (Division of Cardiology), University of Cincinnati College of Medicine, Cincinnati, Ohio. Portions of these studies were supported by the National Institutes of Health, Bethesda, Maryland Program Project Grant HL 22619, Core I and Grant HL 23558 (Dr. Millard) and a grant from the American Heart Association, Southwest Ohio Affiliate (Dr. Fowler). The results of these studies were first reported at the Annual Scientific Meetmg, American Heart Association, November 20, 1980, Miami Beach, Flonda. Manuscript received October 25, 1982; revised manuscript received December 20, 1982, accepted December 23, 1982. Address for reprints: Ronald W. Millard, PhD, Department of Pharmacology and Cell Biophysics, University of Cincinnati Medical Center, 231 Bethesda Avenue, Cincinnati, Ohio 45267. © 1983 by the American College of Cardiology

creased by isoproterenol, hydralazine and the hydralazine-dextran combination. Endocardial flow was increased by amrinone and the combination of hydralazine and dextran. Right ventricular myocardial blood flow increased with all interventions except dextran. Kidney cortical and composite brain blood flows were increased by both dextran alone and by the hydralazine-dextran combinations. Blood flow to small intestine was increased by all interventions as was that to large intestine by all except amrinone and hydralazine. Liver blood flow response was variable. The most pronounced hemodynamic and tissue perfusion improvements during cardiac tamponade were effected by combined vasodilation-blood volume expansion with a hydralazine-dextran combination. Isoproterenol had as dramatic an effect but it was short-lived. Amrinone was the least effective intervention.

a reasonable and rational medical approach needs to become available. The physiologic responses to rapid pericardial fluid accumulation are significant. The first experimental observations were that the arterial pulse rate was rapidly decreased by tamponade and could be improved by intravenous fluid infusion (1-3). Tamponade of the heart caused by increased pericardial fluid results in tachycardia, decreased cardiac output, elevated right and left atrial pressures, a slightly reduced systemic arterial pressure and reduced cardiac stroke volume with marked increase of peripheral vascular resistance. Therefore, it is apparent that temporary improvement of cardiac output and thereby total body perfusion could be produced by several measures, alone or in combination: 1) increasing blood volume, thus filling capacitance vessels with a further increase of central venous pressure to promote ventricular filling and invoke the Frank-Starling mechanism; 2) directly augmenting the cardiac inotropic state by cardiotonic drugs; and 3) administering vasodilator drugs with predominant action on resistance vessels to increase cardiac 0735-1097/83/0601461-10$03.00

1462

J AM COLL CARDIOL 1983,1(6)'1461-70

output while avoiding venous dilation that might further decrease central venous pressure. This study examines the central and peripheral hemodynamic consequences of pharmacologic treatment of acute pericardial tamponade in several groups of dogs. In addition to an untreated control group, there were four groups subjected to different intervention procedures that affected cardiac filling pressure, cardiac inotropic state or peripheral vascular resistance, or a combination of these. The specific interventions employed included intravenous administration of I) isoproterenol for its positive cardiac inotropic effect and to produce peripheral vasodilation; 2) hydralazine to produce vasodilation, followed by administration of dextran to expand circulating blood volume; 3) dextran followed by hydralazine; and 4) amrinone, a newly introduced experimental inotropic agent, to produce increased cardiac inotropism and peripheral vasodilation.

Methods Experimental preparation. Mongrel dogs of either sex weighing between 22 and 34 kg were anesthetized with pentobarbital sodium (25 mg/kg, intravenously) to obtain a surgical plane of anesthesia. Through a left lateral thoracotomy at the fourth intercostal space, catheters were placed in the right atrium and pericardial space. The pericardium was closed snugly and the catheters tunneled to a left lateral subcutaneous pocket. The left atrial and pericardial catheters were filled with heparinized saline solution (0.9%) and their ends closed. The thoracotomy was repaired in four layers and the pneumothorax reduced. Soluble aqueous vitamins and antibiotic agents were administered intraoperatively, followed by daily intramuscular injections of antibiotic agents. Seven days later, the dogs were again anesthetized with pentobarbital sodium (25 mg/kg). A flow-directed catheter inserted into a jugular vein was placed in the pulmonary artery for measurement of cardiac output by thermal dilution, with the proximal injection site in the right atrium to allow recording of cardiac filling pressure. A catheter tip (PC-350, 5F; Millar Instruments) solid state micrornanometer, inserted into a carotid or femoral artery, was located in the basal third of the left ventricular lumen with the aid of fluoroscopy. Tygon catheters were inserted I) in the femoral vein for matched blood infusion during reference arterial sample withdrawal with the microsphere technique and blood volume expansion with dextran-40, and 2) in the femoral artery for blood pressure monitoring by means of a Statham 23Db calibrated electromanometer and for reference arterial sample withdrawal during microsphere injections through the left atrial catheter. Microsphere technique for measuring regional blood flows. Two to four million 15 I-t diameter spheres (3M Company and New England Nuclear) labeled with four different gamma-emitting radionuclides (cerium-141, chro-

MILLARD ET AL

mium-51, strontium-85 and scandium-46) were injected serially at times specified in the following section. Each microsphere moiety was suspended in dextran containing 0.01 % Tween as previously described (4), and thoroughly mixed by vortex rotating alternating with three 30 second periods of ultrasonic agitation. Random microscopic examination of each stock microsphere solution indicated absence of both aggregates and fractured spheres. Microspheres were injected through an indwelling catheter into the left atrium in 2 to 5 ml volumes followed by 10 ml physiologic saline solution to flush the catheter over a 20 second period beginning 15 seconds after the initiation of the 3 minute reference arterial blood sample (nominal withdrawal rate = 10 mllmin). At the same time as the arterial withdrawal. autologous blood, obtained 5 days before the experiment in sterile ACD collection bags, was infused intravenously using a Sarns low flow modular rotary pump (model 6050) at a time, rate and volume equivalent to those of the withdrawal of the arterial sample. The stored blood was converted to heparinized blood by adding 2,000 units heparin and 0.5 mg calcium chloride to each 500 ml of blood before infusion. Cardiac output was determined in triplicate by thermodilution immediately preceding and subsequent to the reference arterial microsphere blood sample. At the end of each protocol, the dog was sacrificed with an overdose of pentobarbital. Tissue samples were taken from left and right ventricles (subepicardial, midmyocardial and subendocardial layers), small intestine (ileum, jejunum and duodenum), large intestine (three colon sites), kidney (cortex and medulla), liver (three major lobes) and brain (cortex, cerebellum, midbrain and medulla oblongata). All tissue samples were weighed to the nearest 10 mg and ranged from 0.75 to 2.25 g. Each tissue was placed in a counting vial and subjected to differential gamma spectroscopy (Packard Instrument Company). Quantitative flows (mllmin per g tissue wet weight) were derived by computer analysis of four simultaneous equations. Determination of microsphere content of each sample was undertaken to ensure statistical reliability of the derived blood flow data according to the criteria of Buckberg et al. (5). Production of acute cardiac tamponade. When hemodynamic variables were stable for at least 30 minutes, microspheres labeled with radionuclide no. I (' 'control" study) were injected. Physiologic saline solution, 37°C, was added to the pericardial space slowly over 10 to 15 minutes through the implanted pericardial catheter until cardiac output was reduced to approximately 50% of the control value. Pericardial fluid volumes were increased by 216 to 288 ml to produce this effect. Microspheres with a second radionuclide label were injected when hemodynamic variables were stable and this reduced cardiac output for at least 15 minutes ("tamponade" study). This was then followed by

BLOOD FLOW DISTRIBUTION IN TAMPONADE

specific interventions in four groups of dogs and with a separate untreated control group.

Interventions Used

=

Isoproterenol group (n 6). An intravenous infusion of a fresh isoproterenol solution (Winthrop Laboratories) (0.5 p,g/kg per min = 3.8 ml/min) was begun after establishment of stable tamponade and continued for 10 minutes . Cardiac output was thereby returned to the baseline control value determined before tamponade . Microspheres labeled with radionuclide no. 3 were injected at this time. The isoproterenol infusion was then terminated while tamponade continued for 15 minutes. Hydralazine-dextran group (n 6). When hemodynamic variables were stable for IS minutes of tamponade, hydralazine, 40 mg, was given as an intravenous injection . Ten minutes later, when a new steady state existed with regard to systemic hemodynamics, microspheres labeled with radionuclide no. 3 were injected . This was followed by blood volume expansion with dextran-40 intravenous infusion. Low molecular weight dextran (6% dextran-75 and 0.9% sodium chloride) (Abbott Laboratories) was buffered by adding 5 ml electrolyte solution* and 3 ml sodium bicarbonate (3.75 g) to each 500 ml. This dextran solution was infused at an initial rate of 100 ml/min and then incrementally reduced to obtain a mean right atrial pressure equal to that attained with tamponade alone before hydralazine . After 15 minutes, when a total of 23 ± 1 ml/kg dextran had been infused, microspheres labeled with radionuclide no. 4 were injected. Dextran-hydralazine group (n 6). After the IS minute tamponade stabilization period dextran-40 was infused at a volume (ml/kg), rate and time identical to that infused in the hydralazine-dextran group (23 ± 1 mllkg body weight). Microspheres labeled with radionuclide no. 3 were then injected . This was followed by an intravenous injection of hydralazine, 40 mg. Ten minutes later, microspheres labeled with radionuclide no. 4 were injected . Amrinone group (n 5). After IS minutes of stable hemodynamic conditions during tamponade, amrinone was injected intravenously in a dose of I mg/kg. At 5 minutes after the amrinone bolus injection, a 10 minute intravenous infusion of 100 p,g/kg per min was begun. Microspheres labeled with radionuclide no. 3 were injected during the 7th minute of this infusion. Tamponade was continued for an additional IS minutes. Untreated control group (n 6). After the IS minute tamponade stabilization period , no further pericardial fluid infusions or withdrawals were made and no intravenous fluids were administered in this group . Microspheres labeled

=

=

=

=

*Electrolyte solution = magnesium chlonde . 6 water = 16.72 g/liter, potas sium chloride = 2974 giliter; potassium = 15.6 mg/rnl; magnesium = 2 mg/ml: calcium = 10 rng/rnl: calcium chloride (anhydrous) = 27 68 g/hter

J AM COLL CARDIOL 1983;1(6):1461 -70

1463

with radionuclides no. 3 and 4 were injected respectively at 25 and 40 minutes of stable cardiac tamponade . These times correspond to those when microspheres were injected in the treated groups. Removal of tamponade. At the end of all experiments , between 92 and 99.5 % of the added pericardial fluid was recovered . Blood gases. All blood gases and blood pH measurements were made with the BMS-3 MK2 blood microsystem (Radiometer, Copenhagen). Arterial blood pH was adjusted by the addition of sodium bicarbonate to maintain a range between 7.30 and 7.48 throughout all experiments. In all groups, arterial blood partial pressure of oxygen (Po 2) exceeded 100 torr before tamponade and partial pressure of carbon dioxide (PC02) ranged from 38 to 49 torr. Adjustments in blood gases were not made during the tamponade or interventions . Arterial hematocrit averaged 40% before tamponade in all groups . Statistical analysis. All results are expressed as mean ± standard error. Data were analyzed using the analysis of variance . Data for each treatment group were compared with data in the untreated group and with preceding data within the same group . Hemodynamic and regional blood flow data were obtained in all animals in each group studied with the exception that no brain samples were taken in the amrinone group. Probability (p) estimates less than 5% were the criteria for rejection of the null hypothesis.

Results Hemodynamic effects of cardiac tamponade (Fig. 1 and 2). The addition of pericardiaI fluid was continued until both pericardial and mean right atrial pressures increased to 10.8 ± 0.8 and 12.0 ± 0.5 mm Hg, respectively, so that cardiac output was reduced by approximately 50%, from 3.53 ± 0.28 to 1.7 ± 0.18 liters/min. Stroke volume accordingly decreased from 18.8 ± 1.3 to 8.4 ± 0.8 ml. This was accompanied by a significant compensatory increase (p < 0.05) in total systemic vascular resistance from 4,110 ± 281 to 6,370 ± 424 dynes-sec -em':". Heart rate increased minimally from 188 ± 4 to 204 ± 9 beats/min. Left ventricular systolic and arterial systolic, diastolic and mean pressures decreased significantly (p < 0.05). Hematocrit increased from 40 to 50% and arterial PC02 decreased from 42 to 33 torr. Consonant with the sharply elevated systemic vascular resistance and decreased cardiac output , blood flow to all measured organs except the liver decreased by 25 to 64%. Left ventricular subepicardial flow decreased from 178 ± 13 to 73 ± 14 ml/min per 100 g. and subendocardial flow decreased from 220 ± 12 to 78 ± II ml/min per 100 g without significant change in the transmural flow distribution ratio; the endocardial/epicardial flow ratio of 1.24 ± 0.03 decreased to I. 12 ± 0.06 . Right ventricular transmural

1464

MILLARD ET AL

J AM COLL CARDIOL 1983. 1(6).1461-70

flow also decreased from 79 ± 9 to 38 ± 6 mllmin per 100 g. Blood flow to the renal cortex decreased from 646 ± 53 to 487 ± 86 mllmin per 100 g as did flow to other visceral organs: small intestine, 71 ± 16 to 21 ± 3 and large intestine, 87 ± 15 to 41 ± 7 ml/min per 100 g. Arterial flow to hepatic tissue was low and variable and was only slightly reduced from 23 ± 5 to 15 ± 5 ml/min per 100 g. Average blood flow to combined samples of brain regions decreased from 43 ± 6 to 29 ± I ml/min per 100g. The following data are for the untreated control group and were mimicked in all other groups before drug treatment . In the untreated control group, a small but gradual spontaneous improvement in all hemodynamic variables occurred during the tamponade period. For example, cardiac output increased 18 ± 8%, from 1.70 ± 0.18 to 2.00 ± 0.21 liters/min during tamponade. Arterial mean pressure increased by IS ± 6%. Pericardial pressure decreased by 1.2 ± 0.2 mm Hg. However, total systemic vascular resistance was unchanged. Organ blood flows were variably but insignificantly changed over initial tamponade values. Isoproterenol group (Fig. 1 and 2). Isoproterenol increased heart rate slightly from 200 ± 15 to 216 ± 12 beats/min and significantly decreased mean arterial blood pressure from 114 ± 6 to 96 ± 3 mm Hg (p < 0.002); stroke volume increased from 10.6 ± 1.7 to 20.7 ± 2.0 ml, as did cardiac output, from 1.80 ± 0.10 to 4.38 ± 0.21 liters/min (p < 0.0005) or 168 ± 14 ml/kg per min. Both pericardial and mean right atrial pressures decreased significantly. Total systemic vascular resistance fell from 4,983 ± 395 to 1,767 ± 89 dynes·s·cm - 5 . Blood flow to right ventricular myocardium, left ventricular subendocardium and subepicardium, small and large intestine and liver was increased significantly as compared with the small spontaneous changes seen in the control group (Fig. 2). The left ventricular transmural flowdistribution ratiodecreased slightly from 1.08 ± 0.08 to 0.93 ± 0.05 , but was significantly reduced compared with the baseline value of 1.21 ± 0.06 (p < 0.05). No significant changes in arterial blood gases, pH or hematocrit were caused by isoproterenol. Hydralazine-dextran group (Fig. 3 and 4). Hydralazine did not affect either heart rate or mean arterial blood pressure (Fig. 3). Both pericardial and mean right atrial pressures were reduced significantly to 10.2 ± 0.4 and 10.3 ± 0.04 mm Hg, respectively. Cardiac output was increased (from 1.56 ± 0.13 to 2.65 ± 0.23 liters/min [p < 0.02] or to 127 ± 12 ml/kg per min) as was stroke volume (from 8.0 ± 0.9 to 12.7 ± 1.1 mI). Total systemic vascular resistance decreased significantly from 7,350 ± 490 to 4,300 ± 392 dynes-s-cm >'. Tissue blood flow was improved to left ventricular subendocardium and subepicardium, right ventricular myocardium and small intestine (Fig. 4) . The group mean changes in blood flow to composite brain tissue, large intestine, kidney and liver were not significantly increased over tamponade values. The left ventricle endocardial/epicardial flow ratio decreased slightly, but insignificantly, from

HEART RATE

250]

(bpm)

p-~ ~

150

MEAN RIGHT 10 ATRIAL PRESSURE (mmHg)

o

20 STROKE VOLUME (ml)

10

6000 SYSTEMIC VASCULAR RESISTANCE

3000 (dynes· sec- em 5 )

BASELINE

I

I 0 I

TAMPONADE Figure 1. Hemodynamic data during cardiac tamponade in the untreated control group (e-e) and in the group treated with isoproterenol (ISO) (0---0). Asterisks indicate values significantly different (p < 0.05) from those of the untreated group at the same point in the protocol. Daggers indicate significant change (p < 0.05) from tamponade values within the same group. Vertical bars represent standard error of mean values.

1.22 ± 0.10 to 1.04 ± 0.06 . Blood gases, pH and hematocrit values were unchanged. The addition of dextran after peripheral vasodilation by hydralazine produced. by design, a restoration of pericardial pressure to levels recorded during tamponade before administration of hydralazine. No change in heart rate or mean arterial blood pressure occurred. Cardiac output was, however, increased further, to 3.93 ± 0.54 liters/min or 206 ± 19 ml/kg per min, as was stroke volume, to 18.8 ± 2.3 ml (p < 0.05 versus hydralazine alone). Total systemic vascular resistance decreased slightly to 3,242 ± 417 dynes-s-cm r ' (Fig. 3). In addition to the hemodynamic effect of dextran after hydralazine treatment, a significant hemodilution resulted, whereby hematocrit was reduced from 47 to 37%.

BLOOD FLOW DISTRIBUTION IN TAMPONADE

1465

J AM COLL CARDIOL 1983.1(6).1461-70

BRAIN

J 200

LV EPICARDIUM

100

200 LV ENDOCARDIUM

100

SMALL INTESTIlE

to 13.0 ± 1.1 and 14.7 ± 0.6 mm Hg, respectively. Mean aortic blood pressure increased slightly to 163 ± 10 mm Hg; stroke volume increased from 8.7 ± 0.9 to 15.6 ± 2.3 ml and cardiac output increased from I. 82 ± O. 19 to 3.2 ± 0.58 liters/min (p < 0.05) or to 127 ± 20 ml/kg per min. Total systemic vascular resistance decreased significantlyfrom6,700 ± 561 to 4,458 ± 546dynes's'cm- 5 (Fig. 5). The improvement in cardiac output and reduction in vascular resistance were paralleled by increases in all tissue blood flows measured, except composite brain samples; however, significant increases compared with values in the untreated group were seen only in left ventricular subepicardium and subendocardium and the large intestine (Fig. 6). The left ventricular endocardial/epicardial flow ratio increased slightly from 1.04 ± 0.06 to 1.15 ± 0.07. Hematocrit was markedly reduced from 46 ± 2 to 33 ± 2% by infusion of dextran. With the addition of hydralazine, the effects on hemodynamic variables and regional organ blood flows were indistinguishable from the data obtained in the previous reversed order treatment group (hydralazine-dextran). In both treatment groups, mean oxygen delivery as determined by

100J 0 Figure 3. Hemodynamic effects of hydralazine followed by dextran (e-e) contrasted with values in the untreated control group (0---0). Values and symbols as in Figure I.

KIDNEY

500

~

.. --i--"--'-'i

~_

HEART RATE

200]

(bpm)

100

Ol-L.--=-:,.......................- - - - - : ! ; = _ BASELINE I I~O TAMPONADE

Figure 2. Blood flow to major organs measured with 15 J.t radionuclide spheres for the untreated control group (e-e) and the isoproterenol-treated (ISO) group (0---0). All flows are expressed as mllmin per 100 g tissue wet weight. Values and symbols as in Figure I; LV = left ventricular.

MEAN RIGHT 1 ATRIAL PRESSURE (mmHg)

0o~

20 STROKE VOLUME (m!)

The combination of dextran with hydralazine (Fig. 4) increased myocardial blood flow to left ventricular subepicardium and subendocardium and to right ventricular myocardium more than two-fold (p < 0.05) versus untreated control group values. The left ventricular endocardial/epicardial flow ratio increased slightly to 1.14 ± 0.12. Somewhat smaller but equally significant increases occurred in blood flow in kidney cortex and small and large intestine. Dextran-hydralazine group (Fig. 5 and 6). When dextran was given first during tamponade, at a similar volume per kilogram body weight, rate of infusion and time course as that in the hydralazine-dextran group, both pericardial and mean right atrial pressures were significantly increased

10

9000 SYSTEMIC VASCULAR RESIST ANCE

6000

(dynes- see .em S) 3000

BASELINE

HYDRALAZINE I

TAMPONADE

DEXTRAN I

1466

J AM COLL CARDIOL

MILLARD ET AL

1983;1(6).1461-70

T

']

BRAIN

~

-,6.~••_••_--!* I

I

300

200 LV EPICARDIUM

ventricular myocardium (+ 159 ± 47%) and small intestine (+ 78 ± 7%) in comparison with the untreated control group values for these tissues. Smaller increases were observed in kidney cortex and large intestine. The left ventricular transmural flow distribution decreased slightly from 1.15 ± 0.13 to 1.00 ± 0.05. No brain samples were taken in this group. Hematocrit was reduced from 44 ± 2 to 38 ± 2% during arnrinone administration, and the calculated oxygen delivery was only slightly increased (Fig. 7).

100

Discussion The ideal pharmacologic agents used in treating cardiac tamponade would improve total cardiac output without inducing tachycardia while maintaining a normal endocardial! epicardial blood flow ratio. Regional blood flow to vital organs (for example, brain, kidney, myocardium, liver and intestine) would be increased or maintained. The present study examines the regional flow changes produced by isoproterenol, a beta-adrenergic agonist that has both positive inotropic and peripheral vasodilator effects; hydralazine, a vasodilator with secondary positive inotropic effects; dex-

30

LV ENDOCARDIUM 200

100

SMALL INTESTINE

100J 0

~ .... ••_.~*.t._ ••__ .u *t I a Figure 5. Hemodynamic changes effected by dextran followed by hydralazine (0---0) contrasted with values in the untreated group (e-e). Values and symbols as in Figure I.

1000

KIDNEY

500

250 HEART RATE

~'-----

(bpm)

BASELINE

g••••

150

HYDRALAZI E DEXTRAN I TAMPONADE I

Figure 4. Regional blood flow during tamponade in untreated control dogs (e-e) contrasted with the distribution obtained with hydralazine followed by dextran (0---0). Values and symbols as in Figure 2.

._--~

i

MEAN RIGHT ATRIAL PRESSURE

10~

(mmHg)

/- .. *t .•••.0:.----

.<5•

o the product of hematocrit and cardiac output was improved despite the hemodilution (Fig. 7). Amrinone group. The administration of amrinone during tamponade increased cardiac output from 1.15 to 0.15 to 1.73 ± 0.21 liters/min (p < 0.05). Total systemic vascular resistance was decreased significantly, from 8,050 ± 728 to 5,180 ± 624 dynes-s-cm (p < 0.05), as were pericardial and mean right atrial pressures, to 8.7 ± 0.8 and 8.9 ± 0.5 mm Hg, respectively. Mean aortic pressure was maintained at 110 ± 6 mm Hg, in contrast to the gradual increase in the untreated control group from 132 ± 10 to 140 ± 7 mm Hg during the same time period. In accord with the vasodilation produced by amrinone, there was significantly increased blood flow to left ventricular subepicardium (+ 216 ± 106%) and subendocardium (+ 148 ± 45%), right ventricular myocardium (+ 148 ± 45%), right t

20 STROKE VOLUME

(mil 10

9000

'

SYSTEMIC VASCULAR RESISTANCE (dynes.see.em

8000 5)

-: {"

,

3000

BASELINE

j

1 "

<,

t 'h"

r " """J'*t

DEXTRAN TAMPONADE

J AM cou, CARDlOl

BLOOD FLOW DISTRIBUTION IN TAMPONADE

1467

1983: 1(6).1461-70

BRAIN

oJ

50

~ 3 •. ·······.o*t -" ""'-"':t •

r

300

LV EPICARDIUM

y/

200

'fl." 1 '. 10

t

.'

c-,.;J;./••/.~_~

300

LV ENDOCARDIUM

200

100

~",.m_ ' ]

~-------'i"", ••••••••'O·t ..

/: :

~*t

KIDNEY

500

~._-~~

Isoproterenol. In previous studies (13), we showed that isoproterenol improves cardiac output in cardiac tamponade. A serial study by Martins et al. (11) with multiple catecholamines in the same dogs with repeated acute tamponade confirmed these observations and the authors extended them to study regional organ perfusion. They observed a tachycardia-dependent reduction in transmural endocardial/epicardial myocardial blood flow ratio. In the present study, isoproterenol in a dose designed to return cardiac output to baseline pretamponade levels increased myocardial blood flow without alteration of heart rate or endocardial/epicardial flow ratio and increased blood flow to small intestine but not to kidney and brain. Martins et al. (11) found that isoproterenol increased flow in myocardium and skeletal muscle but not in brain, kidney or small bowel. Martins et al. (11), however, in studying dogs with chloralose-urethane anesthesia, noted a significant tachycardia caused by isoproterenol that was associated with an apparent adverse distribution of myocardial blood flow such that the endocardium was less well perfused than the epicardium. In our present study, the tachycardia (mean heart rate 188/ min) at rest, attributable to the pentobarbital anesthesia, was not increased significantly by isoproterenol. No changes in

Figure 7. Effect of cardiac tamponade and subsequent treatment on total oxygen delivery (cardiac output [e.O.] times hematocrit [HCL] illustrated by mean group values. A = amrinone group; C = control group; D = dextran group; H = hydralazine group; I = isoproterenol group.

0]~--=-:=-=:-----===:+:-:------:-J~ BASELINE DEXTRAN HYDRALAZINE I TAMPONADE I

Figure 6. Blood flow in ml/min per 100 g tissue wet weight is shown for untreated control group (e-e) and dextran-hydralazine group (0---0). Values and symbols as in Figures I and 2.

200

tran, a plasma volume expander, and amrinone, a proposed new inotropic agent. The combination of dextran and hydralazine was also evaluated.

Effects of Interventions Endocardial/epicardial flow ratio. Total myocardial blood flow, regardless of technique employed, has been shown to decrease in relation to the degree of pericardial tamponade (3,6-11). In contrast to Wechsler et al. (9), other investigators (3,11) failed to note a significant reduction of the endocardial/epicardial flow ratio during tamponade when using 15 JL radionuclide spheres. The findings of our current study confirm those of the latter investigators (3,11); that is, the endocardial/epicardial flow ratio decreased only slightly. The smaller spheres (8 JL) used by Wechsler et al. (9) may be the cause of disparate results due to a greater endocardial microsphere pass-through during tamponade relative to epicardial entrapment (12).

150

C.O.·Hct.

_----C

100

A 50-L.----lIL-----,,;l;----~

;>1

BASELINE

t

;>

INTERVENTION,

TAMPONADE

1468

J AM COLL CARDlOL 1983;1(6) 1461-70

the endocardial/epicardial flow ratio were seen with tamponade alone or when isoproterenol or other intervention was compared with flow distribution during tamponade. Hydralazine. Vasodilators such as hydralazine, having a predominant arteriolar effect, and nitroprusside, with mixed arterial and venous sites of action, have been demonstrated to increase cardiac output during experimental cardiac tamponade (3,14). Our earlier study (14), conducted in dogs anesthetized with morphine-chloralose, showed hydralazine, 40 mg, given during tamponade to effectively increase cardiac output to baseline pretamponade levels. The effects of hydralazine may include reflex augmentation of myocardial inotropic state; however, the slight direct positive inotropic effect of hydralazine has been demonstrated only when high concentrations exist in the coronary circulation (15). In contrast to the study by Gascho et a1. (3), we found that hydralazine increased blood flow to small intestine, large intestine, brain and left ventricular subendocardium and subepicardium. Hydralazine did not reduce the transmural myocardial blood flow distribution ratio (endocardial/ epicardial) in our study, again in contrast to the significant reduction noted by Gascho et a1. (3). The differences may be in part related to the dosages of hydralazine used: 5 to 20 mg by Gascho et a1. (3) and 40 mg in our study. Dextran. Blood volume expansion has been examined as a potentially beneficial intervention in cardiac tamponade by several investigators (1-3). Dextran, saline solution and autologous blood infused during tamponade have generally improved cardiac output presumably through increased cardiac filling pressure and ventricular end-diastolic volume and recruitment of the intrinsic myocardial Frank-Starling mechanism (16). The effects of blood volume expansion with dextran on regional organ perfusion in our study are similar to those recently reported by Gascho et a1. (3), but somewhat less dramatic. Significant further improvements in organ perfusion were produced by blood volume expansion with dextran in animals that had received hydralazine. This protocol (dextran followed by hydralazine) is common to the present study and that of Gascho et a1. (3). Significant increases in blood flow to all major organs except brain occurred with dextran infusion; after the addition of hydralazine there were further increases and, in addition, a significant increase in brain blood flow. Despite hemodilution, dextran infusion combined with hydralazine dramatically increased estimated total oxygen delivery. Regardless of sequence, the combination of volume expansion and hydralazine significantly improved blood flow to myocardium, kidney, brain and large intestine in our study, while Gascho et a1. (3) reported flow increases only to skeletal muscle, kidney and myocardium. These disparities may result from differences in preparation, anesthesia, degree of tamponade, dextran volume added, dose of hydralazine used, or a combination of these. The severity of tamponade produced by Gascho et a1. (3) was considerably

MILLARD ET AL

greater than that we produced, both previously (14) and in the present study. This may have resulted in an enhanced sympathetic state facilitated by the nonbarbiturate anesthesia (17) which opposed the direct vascular action of hydralazine in the study by Gascho et a1. (3). Amrinone. Amrinone was employed in the present study to evaluate its purported beneficial actions in depressed hearts (18,19). In patients with severe congestive heart failure refractory to digitalis therapy, amrinone has been of arguable benefit for improvement of both cardiac output and stroke volume and of renal function. A direct vasodilator action was demonstrated previously by us (20), but only limited positive inotropic action was noted in healthy myocardium. As previously observed with nitroprusside (14), amrinone significantly decreased cardiac filling pressure. Thus, despite a decrease in peripheral vascular resistance (afterload), the decrease in right atrial pressure (preload) probably minimized hemodynamic improvement by amrinone during tamponade. The results obtained with amrinone may indicate some inotropic action (21). A moderate reduction of peripheral vascular resistance was reflected in increased myocardial and gut blood flow. These findings are similar to those previously reported by us (20) on both in vitro vascular smooth muscle strips and in vivo studies of inotropic, chronotropic and vascular resistance changes effected by amrinone. The ability of amrinone to reduce left ventricular enddiastolic pressure and increase cardiac output has been demonstrated in dogs with acute ischemic heart failure (22) and in patients with chronic congestive heart failure (23). The dose used by Jentzer et a1. (22) was identical to that we used in dogs with tamponade. Jentzer et a1. failed to demonstrate significant increases in major organ blood flow. This difference from our findings may relate to the existence of a higher sympathetic tone during tamponade. Amrinone, in the dosage employed in the present study, appears to be of limited value in improving hemodynamic status in cardiac tamponade.

Differences in experimental protocol between this study and previous studies. A marked difference between this study and previous studies (3,11) is the inclusion of a control, untreated group of dogs subjected to cardiac tamponade in which hemodynamic and regional blood flow changes over time were determined. Additionally, the present study was conducted in dogs anesthetized with pentobarbital 1 week after the initial operation when catheters were implanted, while that of Martins et al. (11) was conducted under urethane anesthesia within 10 to 30 minutes after surgery was completed and the thoracotomy repaired. In our study, the use of pentobarbital gave rise to considerable tachycardia in the control period, which tends to minimize further increases in heart rate that might be expected with tamponade and with isoproterenol. Regional distribution of

BLOOD FLOW DISTRIBUTION IN TAMPONADE

J AM COLL CARDIOL

1469

1983.1(6): 1461-70

blood flow may be different after pentobarbital anesthesia when compared with a control state (24). Martins et al. (11) added 150 to 200 ml fluid to the pericardial space in mongrel dogs weighing 18 to 30 kg to increase intrapericardial pressure to 10 mm Hg or to reduce mean arterial blood pressure by at least 30 mm Hg. Gascho et al. (3) added 100 to 280 ml to the pericardial space of mongrel dogs weighing 19 to 31 kg until mean arterial blood pressure fell by 30 to 40 mm Hg, but pericardial pressures were not reported. Cardiac output in those studies was reduced by 70 and 65%, respectively. In the present study, dogs weighing 22 to 34 kg required 216 to 288 ml saline solution added to pericardial space to increase pericardial pressure to 10 to 13 mm Hg so that, by design, cardiac output was depressed by 50%. As a consequence, mean arterial pressure was reduced by 10 to 40 mm Hg. Thus, the present chronically prepared dogs required approximately 10 to 14 ml saline solution/kg body weight added to the pericardial space to accomplish nearly the same hemodynamic end points achieved by Gascho et al. (3) using 5 to 9 ml/kg and Martins et al. (11) using 6 to 8 ml/kg. This difference suggests that 1) pericardial compliance was greater in the present study; 2) the heart was acutely dilated and more easily compressed in the acute tamponade procedure used by the previous investigators (3,11); 3) that these investigators reduced pericardial volume during the surgical procedure; or 4) that myocardial compliance is less in the chronically prepared animals used in the present study. Advantage of the untreated control group. In contrast to many previous studies examining interventions to improve hemodynamics and major organ perfusion in cardiac tamponade, our study included an untreated control group to document time-dependent changes in hemodynamic variables throughout the study protocol. Thus, data from treated groups could be compared not only with internal control values but also with data from the untreated control group to compensate for the small improvement in tamponade dynamics that takes place with the passage of time. Although a gradual improvement in all hemodynamic variables occurred during tamponade in the control group, in no variable was there a significant change. This stability of the animals included here has not been previously demonstrated in acutely prepared tamponade studies. The tendency toward improvement in arterial mean pressure, decrease in pericardial pressure and slight increase of organ blood flow with the passage of time in these acute tamponade preparations might result from several factors. These include small leakage of fluid from the pericardial sac, absorption of fluid from the pericardial sac, decrease in the elevated systemic resistance, an increase in adrenosympathetic cardiac inotropic activity in response to the decrease in blood pressure or pericardial hysteresis. We cannot identify the responsible factors, but the recovery of between 92 and 99% of the added pericardial saline solution at the end of the tamponade makes leakage

or absorption of pericardial fluid an unlikely explanation. Hysteresis of the pericardium may well be a factor, because hysteresis occurs in acute tamponade experiments (25). The results of this study pertain to the animal with barbiturate anesthesia and may not be representative of those produced in human patients with compromised myocardial function or long-standing cardiac tamponade. The hemodynamic responses were collected under conditions in which heart rate was high and was not a significant participating variable. This situation is uncommon in patients with pericardial tamponade in whom changes in heart rate may be expected to participate in the integrated cardiovascular responses to therapeutic interventions (26,27).

Summary Conclusions and Clinical Implications The present study showed that isoproterenol given during tamponade significantly improved blood flow to the small intestine without impairing renal cortical flow or reducing the endocardial/epicardial blood flow ratio. Hydralazine alone increased blood flow to the small intestine without decreasing myocardial endocardial/epicardial blood flow ratio. Hydralazine plus dextran increased blood flow to the brain, epicardium, endocardium, small intestine and renal cortex without significantly decreasing endocardial/epicardial blood flow ratio. Amrinone increased blood flow to endocardium, epicardium and the small intestine, but did not increase cardiac output. Thus, no agent we tested adversely affected endocardial/ epicardial flow ratio during tamponade. Of the single agents tested, dextran had the greatest effect on regional blood flow during cardiac tamponade, significantly increasing flow to the endocardium, epicardium, small intestine and renal cortex. Only the combination of dextran and hydralazine increased regional flow to all areas investigated. No agent decreased regional organ flow when given during tamponade. Implications. We must be cautious in extrapolating the results of these studies to human patients with cardiac tamponade. Regional flow responses to pharmacologic agents may be affected by the degree and duration of tamponade, the anesthesia employed and the dosage of drugs used, as well as by the blood flow volume, inotropic state of the myocardium and the responsiveness of the peripheral vasculature. The dose of isoproterenol used in these studies is much larger than that usually given to patients. Underlying disease states may be important in the response of patients with tamponade to vasoactive drugs. Kerber et al. (27) found that volume expansion with saline solution followed by the vasodilator nitroprusside did not increase cardiac output in patients with cardiac tamponade, unlike the results that had been reported in animal experiments (14). However, their patients, on the average, had an initially low systemic vascular resistance.

1470

MILLARD ET AL

J AM CaLL CARDIOL 1993. 1(6) 1461-70

We appreciate the generous supply of arnrinone by Sterlmg-Winthrop Research Laboraton es. Technical assistance for the conduct of the studies was provided by Robin Smith, Pamela Portner and Cooper Lewis. Manuscript preparation by Janet Simons and Roberta Waltz and Illustrations by Gwen Kraft are sincerely acknowledged.

13.

14.

15.

References I. Cooper FW Jr. Stead EA Jr. Warren JV. Beneficial effect of intra-

venous infusions in acute pericardial tamponade. Ann Surg 1944;120:822-5. 2. Treister B, Gianelly RE. Cohn KE, Harrison DC. The circulatory effects of isoproterenol, acetylstrophanthidin and volume loading in acute pencardial tamponade. Cardiovasc Res 1969;3.299-3 05. 3. Gascho JA, Martins JB, Marcus ML, Kerber RE. Effects of volume expansion and vasodilators in acute pericardialtarnponade AmJ Phys101 1981;232:H49- 53. 4. Millard RW. Baig H, Vatner SF. Cardiovascular effects of radioactive rnicrosphere suspensions and Tween 80 solutions. Am J Physiol 1977;232:H331-4. 5. Buckberg GD. Luck JC, Payne BD. Hoffman HE. Archie JP, Fixler DE Some sources of error in measuring regional blood flow with radioactive microspheres. J Appl Physiol 1971;31.598- 604. 6. O'Rourke RA, Fischer DI', Escobar EE, Bishop VS, Rapaport E. Effect of acute pericardia] tamponade on coronary blood flow. Am J Physiol 1967;212:549-52 . 7 Jarmakani JM. McHale PA, Greenfield JC The effect of cardiac tamponade on coronary hemodynamics In the awake dog Cardiovasc Res 1975 ,9.112-7 . 8. Crystal GJ. Bashour FA. Downey HF, Parker PE. Myocardial blood flow and oxygen consumption during moderate cardiac tamponade: role of reflex vasoconstriction. Proc Soc Exp Bioi Med 1979;160:658 9. Wechsler AS, Auerbach BJ. Graham TC. Sabrston DCJr. Distribution of mtramyocardial blood flow during pericardial tamponade. Correlation with microscopic anatomy and intrinsic myocardialcontractility. J Thorac Cardiovasc Surg 1974;68:847-5 6. 10. Frank MJ. Naduni M. Lesniak LF. Hilmi Kl. Levmson GE. Effects of cardiac tamponade on myocardial performance, blood flow and metabolism. Am J Physiol 1971 :220.179-85. 11. Martins lB. ManuelWJ. Marcus ML, Kerber RE. Comparative effects of inotropic agents on organ perfusion In acute pencardial tamponade. Am J Cardiol 1980;46:59-66 12. Heymann MA. Payne BD. Hoffman HE, Rudolph AM. Blood flow

16. 17.

measurements with radioactive-labeled particles Prog Cardiovasc DIS 1977,20:55-79 Fowler NO. Holmes JC. Hemodynamic effects of isoproterenol and norepinephnne In acute cardiac tamponade. J Clin Invest 1969:48:5027. Fowler NO. Gabel M. Holmes JC. Hemodynamic effects of nitroprusside and hydralazine in experimental cardiac tamponade. CIrculation 1978;57:563-7 Khatri I. Uemura N, Notargiacorno A, Freis ED. Direct and reflex cardiostimulaung effects of hydralazine. Am J Cardiol 1977:40:3842. Starling EH. The Linacre Lecture on the Law of the Heart. London: Longmans, Green. 1918. Michie DD. A plea for evaluating hemodynamic functions In dogs without the use of general anesthetic agents. Am Heart J 1970;80:7145.

18. de Guzman NT. Munoz O. Palmer RF. Davolos D. Alousi A Cluneal evaluation of amrinonc a new inotropic agent (abstr) Circulation 1978:58(suppl 11) .11-1 83. 19. LeJemtel TH. Keung E. Sonnenblick EH. et al. Amrinone. a new nonglycosrdrc, non-adrenergic cardiotonic agent effective in the treatment of Intractable myocardial failure in man. Circulation 1979:59:1098104. 20. Millard RW. Dube G. Grupp G. Grupp I. Alousi A, Schwartz A. Direct vasodilator and positive inotropic actions of arnrinone. J Mol Cell Cardiel 1980;12.647-52 21 . Alousi AA, Farrah AE, Lesher GY, Opalka CJ. Cardiotonic activity of amrinone-WIN 40680 [5-amino-3.4'-blpyridin-6(IH)-one! Cire Res 1979.45:666-77. 22. Jcntzer JH, LeJemtel TH, Sonnenblick EH, Kirk ES. Beneficial effect of arnnnone on myocardial oxygen consumption during acute left ventricular failure in dogs. Am J Cardiel 1981;48:75- 83. 23. Klein NA. Siskind 51. Fnshman WHoSonnenblick EH, LeJemtel TH. Hemodynamic comparison of intravenous amnnone and dobutarru ne In patients With chrome congestive heart failure. Am J Cardiol 1981;48:170-5 . 24. Forsyth RP. Hoffbrand BI. Redistnbunon of cardiac output after sodrum pentobarbital anesthesia in the monkey. Am J I'hYSlOl 1970;21 8:214-7. 25 Morgan BC. Guntheroth SG. Dillard DH. Relationship of pericardial to pleural pressure dunng quret respiration and cardiac tamponade. Circ Res 1965.16:493- 8. 26. Guberman BA. Fowler NO, Engel PJ, Gueron M. Allen JM. Cardiac tamponade In medical patients. Circulation 1981;64:633- 40 27. Kerber RE. Gascho JA, Litchfield R, Wolfson P, Oh D. Pandian NG. Hemodynamic effects of volume expansion and mtroprussidc compared With pericardiocentesis in patients withacute cardiac tamponade. N Engl J Med 1982:307:929- 31.