Effect of atrial fibrillation on atrial blood flow in conscious dogs

Effect of atrial fibrillation on atrial blood flow in conscious dogs

Effect of Atrial Fibrillationon Atrial Blood Flow In ConsciousDogs PHILIP A. McHALE, PhD, JUDITH C. REMBERT, PhD, and JOSEPH C. GREENFIELD, Jr., MD T...

868KB Sizes 3 Downloads 86 Views

Effect of Atrial Fibrillationon Atrial Blood Flow In ConsciousDogs PHILIP A. McHALE, PhD, JUDITH C. REMBERT, PhD, and JOSEPH C. GREENFIELD, Jr., MD

This study was undertaken to examine the independent effects of atrial tachycardia, ventricular tachycardia, and atrial fibrillation (AF) on atrial and ventricular blood flow in conscious, heart-blocked dogs using radioactive microspheres. Atrial blood flow averaged 0.54 f 0.08 ml/min/g during the control period at an atrial rate of 124 beats/mln and a ventricular rate of 90 beats/min. Atrial flow increased to 0.72 f 0.12 ml/min/g during atrial pacing at 236 beats/min, but was not significantly altered by ventricular pacing at 200 beats/mln. AF at a ventricular rate of 90 beats/min resulted In atrial

flow values of 0.91 f 0.08 ml/min/g. The ratio of atrial flow to lefl ventricular flow during AF averaged 1.18 f 0.08. Administration of a maximal vasodilating dose of adenosine during AF further increased atrial flow to 2.18 f 0.16 ml/min/g. Atrial tachycardia or AF did not significantly affect ventricular blood flow. These data indicate (1) that atrial blood flow increases significantly during AF, reaching flow values per gram of tissue comparable to those of the left ventricle, and (2) that this flow is regulated by the metabolic needs of the atrial tissue and does not represent maximal vasodilation.

The observation that coronary blood flow is regulated in direct response to metabolic need has been well established for the ventriculaG as well as the atria13p4 myocardium. In conscious, heart-blocked dogs Neil1 et al3 found that a change in atria1 contraction rate from 92 to 201 beats/min increased atrial blood flow from 0.38 to 0.75 ml/min/g of tissue, thus establishing that atria1 blood flow increases with increasing contraction rate, and is presumably the result of increased metabolic activity. White et al4 observed in anesthetized dogs that atria1 fibrillation (AF), with its more rapid and asynchronous pattern of contraction, increased atria1 blood flow 2- to 3-fold above control values. These investigators also observed that rapid atria1 pacing was a less potent stimulus for increasing atrial blood flow than was AF.4 The present study was designed to assess the independent effects of atrial and ventricular tachycardia and of acute AF on atria1 and ventricular blood flow in a

conscious dog model with complete heart block. This animal model eliminated the hemodynamic changes observed in open-chest preparations, and allowed the changes which normally occur in systemic hemodynamics due to the various interventions to be minimized. To test whether this blood flow increase represented maximal atria1 vasodilation, adenosine was administered during AF. From these data, the relative maximal vasodilating capacity of atria1 and ventricular vascular beds was determined.

Methods Surgical preparation: Eleven adult mongrel dogs, weighing 20 to 28 kg, were anesthetized with sodium thiamylal (25 to 30 mg/kg, intravenously). After endotracheal intubaCon, respiration was maintained with an Emerson model 3-PV respirator. A left thoracotomy was performed through the fifth intercostal space. A heparin-filled polyvinyl chloride catheter (3.0 mm, outside diameter) was introduced into the arch of the aorta by way of the left internal mammary artery. The pericardium was opened and the heart was suspended in a pericardial cradle. A Medtronics model 6917A sutureless

From the Department of Medicine, Division of Cardiology, Duke University Medical Center, and the Veterans Administration Medical Center, Durham, North Carolina. This study was supported in part by Grant HL 18468 from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland, and the Medical Research Service of the Veterans Administration, Durham, North Carolina. Manuscript received August 20, 1982; revised manuscript received November 8, 1982, accepted November 9, 1982. Address for reprints: Joseph C. Greenfield, Jr., MD, Box 3246, Duke University Medical Center, Durham, North Carolina 27710.

unipolar pacing lead and a bipolar pacing lead constructed in our laboratory were implanted on the free wall of the right ventricle. Complete heart block was produced by injection of
June 1983

rate-adjustable

ventricular inhibited pulse generator implanted in a subcutaneous pouch. A second bipolar pacing lead was implanted on the right atria1 appendage. Additional heparin-filled polyvinyl chloride catheters were inserted into the left ventricle through the apex and into the left atrium by way of the left atria1 appendage. The catheters and bipolar pacing electrodes were tunnelled dorsally into a second subcutaneous pouch but were not exteriorized until the day of the study to protect them from damage. The chest was closed, the pneumothorax evacuated, and the animal extubated. The implanted pacemaker was used to maintain heart rate throughout the postoperative period. Experimental procedure: Studies were carried out 7 to 10 days after the initial surgery. All dogs were active and appeared to be in good health, were afebrile, and had hematocrits ranging from 38 to 46%. The dogs were placed on the study table on their right sides, and the laboratory was kept dimly illuminated. The subcutaneous pouch containing the catheters and pacing electrodes was anesthetized with 2% lidocaine infiltration. The catheters and wires were exteriorized through a 1 cm skin incision. The incision was closed in 2 layers around the catheters and wires, then sealed with Aeroplast spray dressing to prevent development of an acute pneumothorax. The ventricular rate was set at 90 beats/min using a Medtronics model 9600A rate controller and was left at that rate for the remainder of the study. The right ventricular bipolar electrode was connected to a Grass model S44 stimulator. The atria1 pacing electrode was connected to a Digitimer model 3290 stimulator. The aortic, left ventricular, and left atria1 catheters were cleared and connected to Statham model P23Db pressure transducers. All pressure measurements were referenced to zero at the mid-chest level. Data were recorded with a Hewlett-Packard model 3955-D magnetic tape recorder and model 7700 eight-channel direct-writing thermal recorder. Microsphere technique: Regional myocardial blood flow and cardiac output were estimated by injecting 8 to 10 I.crn carbonized microspheres labelled with the gamma-emitting radionuclides cerium-141, chromium-51, scandium-46, strontium-85, and niobium-95 into the left atrium. Microspheres were obtained as 1.0 mCi of each nuclide in 10 ml of 10% dextran solution. This stock solution was diluted in 10% dextran solution so that 1.0 ml contained approximately 3,000,000 microspheres. The microspheres were mixed before injection by alternate agitation for at least 15 minutes in an ultrasonic bath and a Vortex agitator. Approximately 1.1 ml of microsphere suspension was drawn up into a preweighed 3 ml syringe, several drops of this suspension were placed into each of 5 preweighed plastic counting vials, and the syringe containing the microsphere suspension was reweighed. Blood flow measurements were made with approximately 1.0 ml of microsphere suspension injected over a 5 second interval into the left atria1 catheter and the catheter was flushed with 10 ml of normal saline solution. A reference sample of arterial blood was collected into 6 plastic counting vials from the aortic catheter at a constant rate of approximately 20 ml/min by a withdrawal pump, beginning simultaneously with the microsphere injection and continuing for 90 seconds. Experimental protocol: The atria1 pacing rate was adjusted to obtain the slowest rate at which the atria could be paced without evidence of any escape beats; the ventricular rate was 90 beats/min. A microsphere injection was performed under these conditions to provide the control data. The atria1 pacing rate then was increased at least 100 beats/min over the rate used during control conditions and another microsphere injection was performed. The atria1 pacing rate was returned to its control value and the ventricular rate was increased to 200 beatslmin by pacing the bipolar right ventricular electrode and thus inhibiting the implanted pacemaker. A microsphere

THE AMERICAN JOURNAL OF CARDIOLOGY

Volume 51

1723

injection was performed under these conditions. Bipolar pacing of the right ventricle was terminated and the ventricular rate returned to 90 beats/min and remained at that rate for the rest of the study. AF was induced by rapid (20 Hz) bipolar stimulation of the right atrial pacing electrode. In 8 dogs, AF was maintained spontaneously after a 5-min period of rapid atria1 pacing; 3 dogs required continuous pacing at 20 Hz to maintain AF. A microsphere injection was made when AF or rapid atria1 pacing had been present for at least 15 minutes. In 6 of the 11 dogs included in this report, an intravenous infusion of adenosine (1 mg/kg/min) was instituted while AF was present. A microsphere injection was performed under these conditions when all hemodynamic variables had been stable for at least 5 minutes. Tissue preparation: The animals were killed at the end of data collection by intravenous administration of 30 mg/kg sodium thiamylal followed by 40 mEq of potassium chloride. The heart was removed and the atria were carefully dissected from the ventricular and venous tissues. The atria were separated into 3 pieces representing the right atria1 free wall, the left atria1 free wall, and the interatrial septum. Each atria1 piece was weighed to the nearest 0.1 g. The left and right ventricles were separated and weighed to the nearest gram. All tissue was fixed in 10% buffered formalin for at least 24 hours. The 3 atrial pieces were cut into multiple small sections each weighing approximately 0.5 g. Care was taken to identify any samples that exhibited tissue damage due to formaldehyde injection and these samples were excluded from the blood flow determinations. Ten transmural samples weighing approximately 1 g were obtained from the free wall of the right ventricle midway between the apex and the pulmonary outflow tract. Each of these transmural samples was divided into an epicardial and an endocardial portion. The left ventricle, including the ventricular septum, was sectioned into 4 transverse rings of approximately equal thickness. The 2 middle rings were divided radially into 7 anatomic sections. Each radial section was further subdivided into 4 transmural layers of approximately equal weight from the epicardium (layer 1) to the endocardium (layer 4). Each tissue sample was weighed to the nearest milligram and was placed in a separate plastic vial for subsequent gamma counting. Calculation of myocardial blood flow and cardiac output: Each tissue sample and the reference blood samples were counted for 10 minutes in a Packard model A5912 gamma spectrophotometer. Counts were accumulated in predetermined energy range windows selected to maximize activity from a given isotope while minimizing spillover from the other isotopes used in the study. Spillover ratios for each isotope into the lower energy windows were determined from locally prepared standards and used to correct the raw tissue counts obtained in each window by the standard stripping technique described by Rudolph and Heymans Total blood flow to each myocardial tissue sample (Q,) in milliliters/ minute was calculated using the formula: Qm = Qr. C,/C,, where Qr is the reference blood sample flow rate (milliliters/ minute), C, is the corrected count activity in the myocardial tissue sample, and C, is the total count activity in the reference blood sample. Total myocardial tissue sample blood flow was divided by the corresponding tissue sample weight and expressed as milliliters/minute/gram of tissue. Cardiac output (Q,,, milliliters/minute) was determined from the relation: QcO= Q, . CJC,, where Qr and C, are as defined previously and Ci is the total amount of activity injected into the left atrium. Ci was determined in the following manner: the 5 plastic counting vials containing several drops of microsphere suspension were reweighed and the weight of microsphere suspension in them was determined by subtraction. The empty syringe from the microsphere injection and the 5 vials were counted in the gamma spectrophotometer

1724

ATRIAL

TABLE

I

BLOOD

Anatomic

FLOW

DURING

ATRIAL

FIBRILLATION

Data

Body weight (kg) Right ventricular weight (g) Left ventricular weight (g) Right atrial free wall weight (g) lnteratrial septum weight (g) Left atrial free wall weight (g) Total atrial weight (g) Left ventricular/body weight ratio (g/kg) Left atrial/body weight ratio (g/kg) Right atrial/right ventricular weight ratio Left atriallleft ventricular weight ratio

24 f l(11) 45f2(11) 111 f 5(11) 7.1 f 0.5 (8) 2.7 f 0.1 (8) 6.2 f 0.5 (8) 16.0 f 0.7 (8) 4.6 f 0.2 (11) 0.27 f 0.02 (8) 0.28 f 0.02 (8) 0.06 f 0.01 (8)

1600 CONTROL q]ATRlAt PACE q ATRIAL FIBRILLATION VENTRICULAR PACE

140-

Values are means f standard error. Numbers in parentheses indicate the number of dogs in which these values were measured. Atrial free wall weights were used in ratio computations. Left ventricular weight includes the interventricular septum.

for 10 minutes each. An estimate of the counts/milligram of microsphere suspension was obtained by dividing the total number of counts in each vial by the weight of the microsphere suspension in that vial. The average counts/milligram for the 5 vials was determined and then multiplied by the difference between the full and empty syringe weights to obtain an estimate of the total number of counts contained in the syringe before injection. The total number of counts remaining in the empty syringe was subtracted from this value to determine Ci, the total number of counts injected into the left atrium. The coefficient of variation of the average counts/milligram of microsphere suspension was <5% for each injection. The percentage of the total number of counts contained in the full syringe that was injected into the left atrium ranged from 85 to 94%. Data analysis: Right atria1 free wall flow values were computed by averaging the data obtained from the individual right atria1 samples. A similar procedure was used to obtain flow values for the left atria1 free wall and the interatrial septum. In addition, a combined atria1 flow value was obtained by computing the average flow for all atrial samples, including the interatrial septum. The endocardial/epicardial blood flow ratio for each left ventricular radial section was computed by dividing the blood flow value obtained for layer 1 into the value for layer 4. Data for all right ventricular samples were averaged to provide the right ventricular blood flow values; similarly, all left ventricular data values were averaged to obtain the left ventricular blood flow values. Thus, each dog contributed to the final analysis a single data value for each anatomic region for each intervention, irrespective of the actual number of atria1 or ventricular tissue samples that were obtained. Atria1 rate, ventricular rate, aortic pressure, and left atria1 pressure were measured from the graphic output obtained at the time of microsphere injection. These data were averages over several respiratory cycles before administration of mi-

TABLE

II

Hemodynamic

124f 5 236 f 4’ 124;

5

Ventricular Rate (beats/min) 90f 90 f 89f 199f

1 1 1 1’

Values are means f standard error of 11 dogs. Significantly different (p <0.05) from control values. + Significantly different (p <0.05) from atrial pace and AF values. l

FfGURE 1. Myocardial blood flow values in ml/min/g f 1 standard error are shown for tissue samples obtained from the indicated anatomic regions during the control period and during the interventions of atrial pacing, AF, and ventricular pacing. Asterisks denote a mean value that was significantly different (p <0.05) from the control value; daggers denote a mean value during AF that was significantly different from the value obtained during atrial pacing.

crospheres and again after completion of reference blood sample collection. Atria1 and ventricular vascular resistance indexes were computed by dividing the mean arterial pressure by the myocardial blood flow values determined by the microsphere technique. Statistical analyses were carried out using multiple paired t tests. When appropriate, the Bonferroni method of adjusting the critical value of t for multiple comparisons was employed.7 Throughout this report, a p value <0.05 was considered to represent a significant difference between 2 mean values. Results Specific anatomic weights and computed weight ratios for the atria and ventricles of the dogs included in this report are shown in Table I. Hemodynamic data obtained in 11 dogs during control conditions and during the interventions of atria1

Data

Atrial Rate (beats/min) Control Atrial pace AF Ventricular pace

RF1 VENTRICLE

RICH1 VEIIIRICLE

Mean Aortic Pressure (mm Hg) 95 92 94 98

f f f f

4 4 4 3+

Left Ventricular End-Diastolic Pressure (mm Hg) 5fl 4&l 3f 1”

Cardiac output (literslmin) 3.0 2.7 2.2 3.2

f f f f

0.2 0.2 0.2’ 0.3

June 1983

pacing, AF, and ventricular pacing are summarized in Table II. Mean aortic pressure was 95 f 4 mm Hg during the control period, and did not change significantly from the control value as a result of the interventions. Cardiac output was not changed from its control value by either atria1 or ventricular pacing; however, a significant 26% decrease in cardiac output was observed during AF. Mean left atria1 pressure was 4 f 1 mm Hg during the control period and during the pacing interventions; a slight but significant decrease in mean left atria1 pressure to 3 f 1 mm Hg was observed during AF. Atrial and ventricular blood flow data obtained in 11 dogs during control, atria1 pacing, AF, and ventricular pacing are shown in Figure 1. Transmural distribution of flow in both the right and left ventricle was not affected by the atria1 pacing interventions; thus, only combined values for ventricular flow are given. During the control period, right atrial flow averaged 0.44 f 0.07 ml/min/g and was significantly lower than the average left atrial flow value of 0.62 f 0.08 ml/min/g. Thus, right atrial flow per gram of tissue was approximately 71% of the corresponding flow in the left atria1 tissue. In the combined atria1 samples, atria1 pacing significantly increased atria1 blood flow by 28%, from 0.54 f 0.08 to 0.72 f 0.12 ml/min/g. AF further increased atrial blood flow to 0.91 f 0.08 ml/min/g which represents a significant 26% increase over the blood flow observed during atria1 pacing at 236 beats/min. This pattern of an increase in atria1 flow with atria1 pacing and a further increase with the onset of AF was observed in the individual samples obtained from the right atrium, left atrium, and interatrial septum. Further analysis of these data revealed~that with AF the dogs fell into 2 distinct groups. In 1 group (6 dogs) there were no discrete atrial contractions that resulted in any change in atria1 pressure; in the second group (5 dogs) discrete atria1 pressure changes ranging from I to 3 mm Hg were generated at a rate of 500 to 800/min. In those dogs in which no discrete pressure changes were observed, the atria1 flow increased an average of 0.47 f 0.06 ml/min/g from the control value. In those dogs in which discrete pressure variation’s were observed, the atria1 flow increased 0.25 f 0.10 ml/min/g from the control value. While the difference in these 2 mean values failed to reach statistical significance (p = 0.08) it strongly suggests that AF without a definite organized atrial contraction induces a greater increase in atria1 flow than occurs when a synchronized rapid atria1 pressure wave is produced. Neither atria1 pacing nor AF significantly changed the blood flow to the right or left ventricular tissue samples from that observed during the control period. A 39% increase in right ventricular and a 55% increase in left ventricular blood flow was observed with ventricular pacing at 200 beatslmin compared with the control value. The left ventricular endocardial/epicardial blood flow ratio was 1.32 f 0.03 during the control period and increased slightly but significantly to 1.47 f 0.06 during ventricular pacing. However, ventricular pacing did not significantly alter blood flow to the atrial tissue samples in any of the regions examined. These changes in atria1 blood flow relative to ventricular blood flow can be examined further by com-

THE AMERICANJOURNALOF CARDIOLOGY Volume 5 1

0 0

q q

1725

CONTROL ATRIAL PACE ATRIA1 Fl8RlltATlON VENTRICULAR PACE

I 40-

120-

“0 F l.OOQ L*I 3 0 z

080-

8 0 060;i; 040-

ozo-

R A&V

LA/tv

!A/tA

FIGURE 2. Myocardial blood flow ratio values f 1 standard error computed between the indicated anatomlc regions during the controt period and during ths interventions of atrial pacing. AF, and ventricular pacing. Asterfeks denote a mean value that was significantly different (p <0.05) from the control value; daggars denote a mean value during AF that was significantly different (p <0.05) from the value obtained during atrial pacing. LA = left atrium; LV = left ventricle: RA = right atrium; RV = right ventricle.

puting several blood flow ratios. Figure 2 shows the right atrial/right ventricular, left atrial/left ventricular, and right atrial/left atria1 blood flow ratios during the control period and during the interventions of atrial pacing, AF, and ventricular pacing. The left atrial/left ventricular blood flow ratio was 0.61 f 0.05 at control, increased to 0.77 f 0.06 with atrial pacing, and increased further to 1.18 f 0.08 with AF. With ventricular pacing the ratio decreased to 0.43 f 0.02. The right atrial/left atrial flow ratio was 0.69 f 0.04 in the control period and did not change significantly with any of the interventions. The hemodynamic effects of a maximal vasodilating dose of adenosine in 6 dogs are shown in Table III. In the presence of AF, adenosine administration decreased

TABLE III

Hemodynamic Etfe&s ot Adenoslne Adminldratlon During Atria1 Flbrlllatlon

Control Mean aortic pressure (mm Hg) Left ventricular end-diastolic pressure (mm Hg) Cardiic outout Iliters/mln)

AF Onlv

AF With Adenosina

96 f 2 3fl

96 f 4 2fl

70 f 3‘ 4fl

3.0 f 0.3

2.1 f 0.3

3.0 f 0.2’

Values are means f standard error of 6 dogs. Significantly different (p <0.05) from AF by paired t test. l

1726

ATRIAL

TABLE IV

BLOW

FLOW

DURING ATRIAL FIBFULLATION

Vascular Resistance Indexes (mm Hg/ml/min)

Combined atria Right ventricle Left ventricle

Control

AF

AF With Adenosine

208 f 27 156 f 21 103 f lot

125 f 11’ 198 f 36 121 f 17

38 f 5+ 31 f 21+ 25 f 8+t

flCONTROL •, ATRIAL

FIBRILLATION

NATRIAL

Fi6RlLLATlON

WITH

ADENOSINE

Values are means f standard error of 6 dogs. Significantly different from control value. t Significantly different from control and AF values. t Significantly different from corresponding combined atria value. l

mean aortic pressure from 96 f 4 to 70 f 3 mm Hg. Cardiac output increased 45% from its value during AF alone and reached a value (3.0 f 0.2 liters/min) that was not significantly different from the cardiac output observed during the control period. Left atria1 pressure was not altered by adenosine administration. Atrial and ventricular blood flows during the control period and during AF with and without adenosine are shown in Figure 3. With adenosine administration, atria1 blood flow increased 146% from the value observed during AF alone. This atria1 blood flow value of 2.18 f 0.16 ml/ min/g represents a 303% increase over the corresponding control value. For comparison, left ventricular blood flow increased from a control value of 1.02 f 0.11 to 4.09 f 0.27 ml/min/g with adenosine administration during AF. Thus, adenosine administration resulted in a 300% increase in ventricular blood flow. Vascular resistance indexes computed for the control period and during AF with and without adenosine are shown in Table IV. Administration of adenosine during AF resulted in a 69 f 4% decrease in the atrial resistance index while the right and left ventricular resistance indexes decreased 85 f 4% and 80 f 3%, respectively. This 69% atrial resistance index decrease is significantly lower than the 85% and m decrease in the ventricular vascular resistance indexes. Note that the vascular resistance index for the left ventricle is significantly less than for the combined atria during both the control period and during vasodilation secondary to AF and adenosine administration. Discussion This study was carried out in conscious dogs to avoid the effects of acute surgical trauma and thus obtain the data in as physiologic a preparation as possible. Complete heart block was created so that a constant ventricular rate of 90 beata/min could be maintained during atrial pacing and AF, keeping the perfusion pressure of the atria1 and ventricular vascular beds constant. As shown in Tables II and III, mean aortic pressure remained within a range of 6 mm Hg, except when adenosine was administered. During the control period, atria1 pacing at 124 beats/min resulted in atrial contractions that generated atrial pressure increases ranging from 2 to 12 mm Hg in all dogs. Atrial pacing at 236 beats/min also resulted in discrete atrial contractions that generated comparable pressure increases. During AF, the group of 5 dogs which exhibited discrete atria1 pressure changes did not increase their atria1 blood flow as much as the group of 6

COYNlNfO

RI

AlNlA

YENlRlCLl

YENIRICLE

FKWRE 3. Myocardiil blood flow values in ml/min/g f 1 standard error are shown for tissue samples obtained from the indicated anatomic regions during the control period and during AF with and without adenosine. Aaterlsks denote a mean value obtained during AF with adenosine administration that was significantly different (p <0.05) from the corresponding value obtained during AF alone. Note the change in scale from Figure 1.

dogs which exhibited no discrete atria1 pressure changes. Thus, it seems fair to conclude that atrial blood flow is determined, at least in part, by the frequency of atria1 contraction. In addition, Neil1 et al3 observed in heart-blocked dogs that atria1 vasomotion is independent of ventricular vasomotion, as demonstrated by the fact that atria1 blood flow did not increase in response to ventricular pacing. This observation was confirmed in the present study. The increased flow observed during AF could have resulted from an acute increase in atria1 size, resulting in an increase in atria1 wall stress. Atria1 geometry was not measured in the studies included in this report. However, mean left atria1 pressure was slightly decreased during AF and, unless there is a marked decrease in atrial compliance, it is unlikely that the atrium was larger. Henry et a1,8 using echocardiographic determinations of left atria1 size in human subjects, observed no change in atria1 size before and after an episode of paroxysmal AF. In the present experiment it was not possible to obtain samples of atrial venous blood, therefore, we cannot directly address the question of whether the increased flow observed during AF represents a true increase in metabolic demand. However, White et a1,4using a microspectrophotometric method in anesthetized dogs,

June 1983

observed that the atria1 02 extraction averaged 61% in their experiment when sinus rhythm was present and was not altered by AF. Thus, it is reasonable to assume in our dogs that the metabolic demands of the atria1 muscle are significantly increased during AF and that these demands are met primarily in increasing blood flow. In a study which compared ventricular blood flow in the empty beating heart to the empty fibrillating heart, Kleinman and Wechslerg observed that an 82% increase in flow occurred when ventricular fibrillation was induced. This change is comparable to the 69% increase in atria1 blood flow observed during AF. It is important to know if the elevated blood flow observed during AF represents maximal atria1 vascular vasodilation or if this vascular bed is simply regulating its blood flow at a level consistent with its metabolic demand. In the present study, administration of a maximal vasodilating dose of adenosine during AF resulted in a 146% increase in atria1 blood flow from that observed during AF alone. This observation would indicate that atria1 blood flow during AF is not maximal and thus appears to be under active vasomotor control. The data obtained in our studies and in the study of White et al4 clearly demonstrate that the acute onset of AF induced by electrical stimulation is associated with a significant increase in atria1 flow. This does not answer the question of whether chronic AF is associated with a similar increase in atria1 blood flow and, hence, it is somewhat premature to speculate on the clinical importance of these measurements. On the other hand, if the chronically fibrillating atria can be shown to maintain a high metabolic need per gram of tissue, especially in the hypertrophied atrium, then the presence of a flow-limiting stenosis proximal to the branching of the atria1 arteries may create a hemody-

THE AMERICAN JOURNAL OF CARDIOLOGY

Volume 5 1

1727

namic impairment that has not been appreciated previously. That is, the increase demand for flow in the atria may act to further reduce the already compromised amount of flow available for ventricular perfusion, and hence accelerate symptomatic coronary artery disease. Acknowledgment: We express our gratitude to Kirby Cooper and Eric Fields for surgical preparations, Donald G. Powell and the staff of the Durham Veterans Administration Medical Center Medical Media Production Service, J. Michael Taylor and the staff of the Durham Veterans Administration Medical Center Animal Care Facility, and Dianne Higginbotham for skillful preparation of the manuscript. We are especially grateful to Leigh Stubbs of Medtronics for assistance with the pacing technique and t.o Robert Murdoch for developing the techniques necessary for determining cardiac output from the radioactive microsphere injections.

References 1.

2.

3. 4.

5. 6. 7. 6. 9.

Katz LN, Feinberg H. The relation of cardiac effort to myocardial oxygen consumotion and coronary flow. Circ Res 1958:6:656-669. Miller W’L, Belardinelli L, Dacchus A, Foley DH, kubio R, Beme RM. Canine myocardial adenosine and lactate production, oxygen consumption, and coronary blood flow during stellate ganglia stimulation. Circ Res 1979;45: 708-718. Neil1 WA, Sewell D, Gopal M, Oxendine J. Painter L. Independent reaulation of atrial coronary blood flow by atrial contraction rate in conscioui dogs. Pfluaers Arch 1980:388:193-195. Wht&eCW, Kerber HE, Weiss HR. Marcus ML. The effects of atrial fibrillation on atrial pressure-volume and flow relationships. Circ Res 1982;51:205215. Steiner C, Kovallk Al. A simple technique for productionof chronic complete heart block in dogs. J Appl Physiol 1968;25:631-632. Rudolph AM, Heyman MA. The circulation of the fetus in utero: methods for studyingdistributionof blood flow, cardiac output, and organ blood flow. Circ Res 1967;21:163-184. Walienstein S, Zucker CL, Fleiss JL. Some statistical methods useful in circulation research. Circ Res 1980;47:1-9. Henry WL. Morganroth J, Pearlman AS, Clark CE, Redwood DR ltscoftz SB, Epstein SE. Relation between echocardiographically determined lefl atrial size and atrial fibrillation. Circulation 1976;53:273-279. Kleinman LH, Wechsler AS. Pressure-flow characteristics of the coronary collateral circulation during cardiopulmonary bypass: effects of ventricular fibrillation. Circulation 1978;58:233-239.