- Email: [email protected]
Experimental coronary artery fistula
Experimental coronary artery fistula Despite numerous clinical reports of coronary artery fistulas, there have been no laboratory models. A circumflex coronary-pulmonary artery fistula was constructed with a vein graft in 11 adult foxhounds. A temporary intravascular shunt obviated the need for cardiopulmonary bypass. Measurements of aortic flow, arterial pressure, left ventricular pressure and its derivative, heart rate, and flow in the proximal coronary, distal coronary, and fistula were made with the fistula open and closed in 9 animals. Mean fistula flow was 89 c.c. per minute, representing a 1.1:1 left-to-right shunt. Mean proximal coronary flow increased 211 per cent, and there was a relative steal of 26 per cent of distal coronary flow with the fistula functioning. Phasic flow patterns showed continuous systolic and diastolic flow in the proximal coronary artery and fistula. Despite the striking changes in coronary flow patterns, there was no significant effect on measured left ventricular function. Further uses for this model and variations of it are suggested.
Bruce A. Reitz, M.D., Lynn H. Harrison, Jr., M.D., and Lawrence L. Michaelis, M.D., Bethesda, Md.
V_^ongenital and traumatic coronary artery fistulas are well-recognized entities. More than 200 clinical descriptions have been reported, with documentation of the hemodynamic and angiographic findings and the results of operative treatment. 110 The majority of congenital fistulas (60 per cent) originate in the right coronary artery, and about 90 per cent of all fistulas terminate in the right side of the heart or pulmonary artery.1' -'■ lil " The magnitude of the left-toright shunt varies from not being measurable up to 2.2:1, with most being about 1.1:1. 3 Most patients are asymptomatic, but others have had congestive heart failure (14 per cent), angina pectoris (7 per cent); exertional dyspnea (28 per cent), or bacterial endocarditis (6 per cent). 11 The presence and severity of symptoms presumably are related to the size of the shunt as well as to a possible steal phenomenon in the distal coronary circulation,--x although this latter mechanism has never been demonstrated directly.12 Despite the clinical interest in coronary From the Clinic of Surgery, National Heart and Lung Institute, National Institutes of Health, Bethesda, Md. 20014. Received for publication March 13, 1974.
278
artery fistulas, there have been no animal models for laboratory study. In our experiments, a model of left circumflex coronary artery-to-pulmonary artery fistula was created in dogs. The flow patterns in the coronary artery and fistula were recorded, and left ventricular function was assessed with the fistula open and closed. Methods Operative technique. Eleven adult foxhounds ranging in weight from 17 to 24 kilograms were anesthetized with intravenous sodium thiamylal (25 mg. per kilogram) and halothane (0.5 to 1 per cent). Ventilation with 50 per cent oxygen was maintained through a cuffed endotracheal tube and Harvard respirator. A left thoracotomy was made through the fourth intercostal space, the left atrial appendage was retracted superiorly, and the proximal 2 cm. of the circumflex coronary artery was dissected free. A stiff polyvinyl catheter (Cobe Laboratories, Inc., Denver, Colo.) was placed in the left carotid artery and connected by way of a three-way stopcock to a 10 inch segment of a smaller (16 gauge) Teflon catheter (Fig. 1). A segment approximately 5 cm. in length
Volume 69
Experimental coronary artery fistula 2 7 9
Number 2 February. 1975
PERFUSER CAROTID ARTERY
Ao
FLOW
SITE OF PA ANASTOMOSIS
•
FLOW
PROBE
PLACEMENT
VEIN SEGMENT
CIRCUMFLEX
CONSTRICTORS
A.
B.
Fig. 1. A, Placement of the intra vascular shunt from the carotid artery through the vein segment and into the distal circumflex coronary artery. This allows anastomosis of the vein to the coronary artery. B, Experimental setup after the fistula is completed. Aortic and left ventricular pressures are measured, and flow transducers are placed on the aorta, coronary artery, and fistula, as shown. Ao, Aorta. PA, Pulmonary artery. LAD, Left anterior descending coronary artery. LA, Left atrium. LV, Left ventricle. AP, Aortic pressure. LVP, Left ventricular pressure.
was taken from the lateral vein of the hind leg or the cephalic vein of the foreleg. This vein segment was then threaded onto the Teflon catheter as shown in Fig. 1, A, so that the venous valves would permit unobstructed flow from the coronary to the pulmonary artery. Sodium heparin (3 mg. per kilogram) was administered, and two 3-0 Tevdek ligatures with small polyethylene keepers were placed around the coronary artery as shown. A 7 to 9 mm. arteriotomy was made, and the Teflon catheter was passed into the distal coronary artery after the distal tie was momentarily relaxed. Perfusion was then begun by turning on the three-way stopcock. Ischemic time was less than 45 seconds, and ventricular fibrillation did not occur during any procedure. The vein segment was anastomosed to the coronary artery with continuous sutures of 6-0 Tevdek. When this was completed, the two ties around the coronary artery were removed, the perfuser pulled back through
the vein, and a bulldog clamp placed on the vein. A side-biting clamp was then placed on the pulmonary artery, and a 4 to 5 mm. button of tissue was removed. The distal end of the vein segment was then anastomosed to the pulmonary artery with 6-0 Tevdek. At the completion of the procedure, a continuous thrill could be felt in the vein segment. Experimental design. After construction of the fistula, the dog was instrumented for hemodynamic study. Left ventricular and aortic pressures were recorded with Statham P23Db transducers. The transducer of an electromagnetic flowmeter (Biotronex Laboratory, Inc., Silver Spring, Md.) was placed around the aortic root. A second, smaller (2 to 3 mm.) flow transducer was placed sequentially on the vein segment, proximal coronary artery, and distal coronary artery as shown in Fig. 1, B. Recordings were made on a multichannel oscillo-
The Journal of
280
Reitz, Harrison, Michaelis
Thoracic and Cardiovascular Surgery
Table I* Parameter
Fistula closed
Heart rate (beats/min.) Arterial pressure (mm. Hg) Left ventricular pressure (mm. H g )
116 97/71 106/3
Left ventricular d p / d t (mm. Hg/sec.) Stroke volume (c.c.) Cardiac output (c.c./min.) Fistula flow (c.c./min.) Proximal coronary (c.c./min.) Distal coronary (c.c./min.) Arterial pH Arterial Po 2 (mm. Hg) Arterial Pco 2 (mm. Hg) Hematocrit (per cent)
1,660 8.8 1,000 0 31 28 7.53 396 28 40.5
± ±
+
+ +
± +
± + +
± ± + +
15 3.3 4.1 5 0.8 120 0.8 65 4 3 .03 48 3 0.8
Fistula
open
Significance
115 ± 16 ± 3 96/71 ± 3 ?
N.S.
106/3 \ J_8
N.S.
1,640 9.0 1,024 89 97 21
N.S. N.S. N.S.
± ± ± ± ± ±
115 0.8 79 9 12 3
N.S.
-
p < p <
0.001 0.02
•Hemodynamic and metabolic data from 9 dogs with a circumflex coronary-pulmonary artery fistula. Values are mean ± S.E., and all 9 dogs are included for each value.
graph (Brush Instruments, Cleveland, Ohio) with a paper speed from 10 to 100 mm. per second. After a 30 minute equilibration period, blood gases and hematocrit values were checked. Nine dogs were satisfactory for study at this point. All variables were measured after the fistula had been open for 2 minutes and again 2 minutes after it had been closed with a clamp. These measurements were repeated in each animal as the small flowmeter was moved from the vein segment to the proximal coronary, distal coronary, and back to the vein segment. Calculations. The measurements of arterial pressure, left ventricular pressure, and heart rate were taken from the oscillographic record. The derivative of left ventricular pressure was electronically figured, and the stroke volume was calculated from the area beneath the aortic flow-velocity signal. All flowmeters were calibrated before and after the experimental series with good agreement. Student's t test was used for determination of statistical significance. Results The data from all nine experiments are summarized in Table I. Coronary flow patterns. Flow measured in the fistulas ranged from 55 to 128 c.c. per minute and averaged 89 ± 9 (S.E.) c.c.
per minute. This represented a left-to-right shunt of 1.1:1, since the mean cardiac output was 1,024 ± 79 c.c. per minute. Under control conditions the flow in the proximal circumflex coronary artery was 31 ± 4 c.c. per minute; it increased abruptly to 97 ± 12 c.c. per minute when the circumflex coronary-pulmonary artery fistula was opened. Concurrently, the measured flow in the distal circumflex decreased from 28 + 3 to 21 ± 3 c.c. per minute, a relative steal of 26 per cent of the coronary perfusion in this artery. Phasic flow patterns are shown in Fig. 2. The flow pattern in the proximal coronary artery changed abruptly from the usual diastolic flow dominance to a continuous systolic and diastolic flow pattern very similar to the pattern recorded in the fistula itself. In the distal coronary artery there was a slight reduction in diastolic blood flow; mean flow signals obtained from another experiment (shown in Fig. 3) demonstrate the definite reduction in perfusion of the distal coronary bed. Left ventricular function. The hemodynamic data are summarized in Table I. Arterial pressure, left ventricular pressure and its derivative, stroke volume, heart rate, and cardiac output did not vary with the fistula open or closed. The coronary blood flow in the proximal coronary artery was
Volume 69 Number 2 February, 1975
Experimental
coronary
C-PAF CLOSED
PROXIMAL CORONARY (CC/min)
DISTAL CORONARY
(CC/min)
FISTULA (CC/min)
2 8 1
C-PAF OPEN
"^^^U4-uu^
ECG
artery fistula
^u*4-*
250 r 125 i-
250125 i0'
25C1250-J
EXP IR035
L SECONDS
Fig. 2. Recordings of hemodynamic parameters measured with the circumflex-pulmonary artery fistula (C-PAF) closed and after it had been opened for 2 minutes. ECG, Electrocardiogram.
MEAN DISTAL CORONARY FLOW (CC/min)
FISTULA OPEN
FISTULA CLOSED
250 r125 0
ECG
i?fi?N^5»f^i^^^ I
I
l _
i
EXP. 1032
i
i
i
i
i
i
i
i
i
i
i
i
SECONDS
Fig. 3. Mean distal coronary artery flow is shown as the fistula is opened for 5 seconds and then closed. ECG, Electrocardiogram. significantly increased (p < 0.001) and the flow in the distal coronary significantly decreased ( p < 0.02) with the shunt open. Discussion In these experiments, a coronary artery fistula was constructed to enter the pulmonary artery. It appears that the same method could be utilized to produce coronary arterial fistulas entering the right atrium, right ventricle, or left atrium. The
vein segments used ranged from 3 to 5 mm. in diameter and produced an average left-toright shunt of 1.1:1, a shunt comparable in magnitude to that measured in most asymptomatic patients." We presume that, with larger vein segments (as from the external jugular vein), larger shunts could be produced. It may be possible to examine the longterm effects of a coronary artery fistula with this model. The tendency for the proximal
The Journal of
2 82
Reitz, Harrison, Michaelis
Thoracic and Cardiovascular Surgery
coronary artery to dilate under conditions of increased flow,1 the chronic hemodynamic and morphologic effects of a fistula, and the natural history of such lesions are important areas to examine. It has been suggested that the greatly increased flow in the proximal coronary artery may contribute to accelerated atherogenesis.■'' If so, another use for the model would be to study this phenomenon. The presence of a steal phenomenon has been inferred from the characteristics of peripheral arteriovenous fistulas1^ and from angiography1' but has not been demonstrated directly. In these experiments with a small circumflex coronary-pulmonary artery fistula, there was a significant decrease (26 per cent) in distal coronary flow. This represents a relative reduction and not a reversal of flow, although flow reversal might be demonstrated with larger shunts. Under the conditions of this experiment, there was no demonstrable change in left ventricular function. This corresponds to the clinical situation in the majority of patients who are asymptomatic and who have normal catheterization findings except for the presence of a left-to-right shunt.' 7 - "' However, the effect of anesthesia, thoracotomy, and the acute nature of the experiments does not warrant the conclusion that the demonstrated decrease in distal coronary flow is without significant effect. REFERENCES 1 Sabbagh, A. H., Schocket, L. I., Griffin, T., Anderson, R. M., Goldberg, S., Fritz, J. M., and O'Hare, J.: Congenital Coronary Artery Fistula,
J.
794, 1973.
THORAC.
CARDIOVASC.
SURG.
66:
2 Parker, F. B., Neville, J. F., Johnson, L. W„ Scrivani, J. V., and Webb, W. R.: Congenital Coronary Artery Fistula From Supernumerary Coronary
Artery,
SURG. 65: 569,
J.
THORAC.
CARDIOVASC.
1973.
3 Jaffe, R. B., Glancy, D. L., Epstein, S. E., Brown, B. G., and Morrow, A. G.: Coronary Arterial-Right Heart Fistulae: Long Term Observations in Seven Patients, Circulation 47: 133, 1973. 4 Bravo, A. J., Glancy, D. L., Epstein, S. E., and Morrow, A. G.: Traumatic Coronary Arteriovenous Fistula, Am. J. Cardiol. 27: 673, 1971. 5 Edis, A. J., Schattenberg, T. T., Feldt, R. H., and Danielson, G. K.: Congenital Coronary Artery Fistula: Surgical Considerations and Results of Operation, Mayo Clin. Proc. 47: 567, 1972. 6 De Nef, J. J. E., Varghese, P. J„ and Losekoot, G.: Congenital Coronary Artery Fistula: Analysis of 17 Cases, Br. Heart J. 33: 857, 1971. 7 McNamara, J. J., and Gross, R. E.: Congenital Coronary Artery Fistula, Surgery 65: 59, 1969. 8 Nora, J. J., McNamara, D. G., Hallman, G. L., Sommervilk, R. J., and Cooley, D. A.: Medical and Surgical Management of Anomalous Origin of the Left Coronary Artery From the Pulmonary Artery, Pediatrics 42: 405, 1968. 9 Sabiston, D. C , Jr., Ross, R. S., Criley, J. M., Gaertner, R. A., Neill, C. A., and Taussig, H. B.: Surgical Management of Congenital Lesions of the Coronary Circulation, Ann. Surg. 157: 908. 1963. 10 Upshaw, C. B.: Congenital Coronary Arteriovenous Fistula: Report of a Case With an Analysis of Seventy-three Reported Cases, Am. Heart J. 63: 399, 1962. 11 Oldham, H. N., Ebert, P. A.. Young, W. G., and Sabiston, D. C , Jr.: Surgical Management of Congenital Coronary Artery Fistula, Ann. Thorac. Surg. 12: 503, 1971. 12 Rowe, G. G.: Inequalities of Myocardial Perfusion in Coronary Artery Disease ("Coronary Steal"), Circulation 42: 193, 1970.
Bruce A. Reitz, M.D., Lynn H. Harrison, Jr., M.D., and Lawrence L. Michaelis, M.D., Bethesda, Md.
V_^ongenital and traumatic coronary artery fistulas are well-recognized entities. More than 200 clinical descriptions have been reported, with documentation of the hemodynamic and angiographic findings and the results of operative treatment. 110 The majority of congenital fistulas (60 per cent) originate in the right coronary artery, and about 90 per cent of all fistulas terminate in the right side of the heart or pulmonary artery.1' -'■ lil " The magnitude of the left-toright shunt varies from not being measurable up to 2.2:1, with most being about 1.1:1. 3 Most patients are asymptomatic, but others have had congestive heart failure (14 per cent), angina pectoris (7 per cent); exertional dyspnea (28 per cent), or bacterial endocarditis (6 per cent). 11 The presence and severity of symptoms presumably are related to the size of the shunt as well as to a possible steal phenomenon in the distal coronary circulation,--x although this latter mechanism has never been demonstrated directly.12 Despite the clinical interest in coronary From the Clinic of Surgery, National Heart and Lung Institute, National Institutes of Health, Bethesda, Md. 20014. Received for publication March 13, 1974.
278
artery fistulas, there have been no animal models for laboratory study. In our experiments, a model of left circumflex coronary artery-to-pulmonary artery fistula was created in dogs. The flow patterns in the coronary artery and fistula were recorded, and left ventricular function was assessed with the fistula open and closed. Methods Operative technique. Eleven adult foxhounds ranging in weight from 17 to 24 kilograms were anesthetized with intravenous sodium thiamylal (25 mg. per kilogram) and halothane (0.5 to 1 per cent). Ventilation with 50 per cent oxygen was maintained through a cuffed endotracheal tube and Harvard respirator. A left thoracotomy was made through the fourth intercostal space, the left atrial appendage was retracted superiorly, and the proximal 2 cm. of the circumflex coronary artery was dissected free. A stiff polyvinyl catheter (Cobe Laboratories, Inc., Denver, Colo.) was placed in the left carotid artery and connected by way of a three-way stopcock to a 10 inch segment of a smaller (16 gauge) Teflon catheter (Fig. 1). A segment approximately 5 cm. in length
Volume 69
Experimental coronary artery fistula 2 7 9
Number 2 February. 1975
PERFUSER CAROTID ARTERY
Ao
FLOW
SITE OF PA ANASTOMOSIS
•
FLOW
PROBE
PLACEMENT
VEIN SEGMENT
CIRCUMFLEX
CONSTRICTORS
A.
B.
Fig. 1. A, Placement of the intra vascular shunt from the carotid artery through the vein segment and into the distal circumflex coronary artery. This allows anastomosis of the vein to the coronary artery. B, Experimental setup after the fistula is completed. Aortic and left ventricular pressures are measured, and flow transducers are placed on the aorta, coronary artery, and fistula, as shown. Ao, Aorta. PA, Pulmonary artery. LAD, Left anterior descending coronary artery. LA, Left atrium. LV, Left ventricle. AP, Aortic pressure. LVP, Left ventricular pressure.
was taken from the lateral vein of the hind leg or the cephalic vein of the foreleg. This vein segment was then threaded onto the Teflon catheter as shown in Fig. 1, A, so that the venous valves would permit unobstructed flow from the coronary to the pulmonary artery. Sodium heparin (3 mg. per kilogram) was administered, and two 3-0 Tevdek ligatures with small polyethylene keepers were placed around the coronary artery as shown. A 7 to 9 mm. arteriotomy was made, and the Teflon catheter was passed into the distal coronary artery after the distal tie was momentarily relaxed. Perfusion was then begun by turning on the three-way stopcock. Ischemic time was less than 45 seconds, and ventricular fibrillation did not occur during any procedure. The vein segment was anastomosed to the coronary artery with continuous sutures of 6-0 Tevdek. When this was completed, the two ties around the coronary artery were removed, the perfuser pulled back through
the vein, and a bulldog clamp placed on the vein. A side-biting clamp was then placed on the pulmonary artery, and a 4 to 5 mm. button of tissue was removed. The distal end of the vein segment was then anastomosed to the pulmonary artery with 6-0 Tevdek. At the completion of the procedure, a continuous thrill could be felt in the vein segment. Experimental design. After construction of the fistula, the dog was instrumented for hemodynamic study. Left ventricular and aortic pressures were recorded with Statham P23Db transducers. The transducer of an electromagnetic flowmeter (Biotronex Laboratory, Inc., Silver Spring, Md.) was placed around the aortic root. A second, smaller (2 to 3 mm.) flow transducer was placed sequentially on the vein segment, proximal coronary artery, and distal coronary artery as shown in Fig. 1, B. Recordings were made on a multichannel oscillo-
The Journal of
280
Reitz, Harrison, Michaelis
Thoracic and Cardiovascular Surgery
Table I* Parameter
Fistula closed
Heart rate (beats/min.) Arterial pressure (mm. Hg) Left ventricular pressure (mm. H g )
116 97/71 106/3
Left ventricular d p / d t (mm. Hg/sec.) Stroke volume (c.c.) Cardiac output (c.c./min.) Fistula flow (c.c./min.) Proximal coronary (c.c./min.) Distal coronary (c.c./min.) Arterial pH Arterial Po 2 (mm. Hg) Arterial Pco 2 (mm. Hg) Hematocrit (per cent)
1,660 8.8 1,000 0 31 28 7.53 396 28 40.5
± ±
+
+ +
± +
± + +
± ± + +
15 3.3 4.1 5 0.8 120 0.8 65 4 3 .03 48 3 0.8
Fistula
open
Significance
115 ± 16 ± 3 96/71 ± 3 ?
N.S.
106/3 \ J_8
N.S.
1,640 9.0 1,024 89 97 21
N.S. N.S. N.S.
± ± ± ± ± ±
115 0.8 79 9 12 3
N.S.
-
p < p <
0.001 0.02
•Hemodynamic and metabolic data from 9 dogs with a circumflex coronary-pulmonary artery fistula. Values are mean ± S.E., and all 9 dogs are included for each value.
graph (Brush Instruments, Cleveland, Ohio) with a paper speed from 10 to 100 mm. per second. After a 30 minute equilibration period, blood gases and hematocrit values were checked. Nine dogs were satisfactory for study at this point. All variables were measured after the fistula had been open for 2 minutes and again 2 minutes after it had been closed with a clamp. These measurements were repeated in each animal as the small flowmeter was moved from the vein segment to the proximal coronary, distal coronary, and back to the vein segment. Calculations. The measurements of arterial pressure, left ventricular pressure, and heart rate were taken from the oscillographic record. The derivative of left ventricular pressure was electronically figured, and the stroke volume was calculated from the area beneath the aortic flow-velocity signal. All flowmeters were calibrated before and after the experimental series with good agreement. Student's t test was used for determination of statistical significance. Results The data from all nine experiments are summarized in Table I. Coronary flow patterns. Flow measured in the fistulas ranged from 55 to 128 c.c. per minute and averaged 89 ± 9 (S.E.) c.c.
per minute. This represented a left-to-right shunt of 1.1:1, since the mean cardiac output was 1,024 ± 79 c.c. per minute. Under control conditions the flow in the proximal circumflex coronary artery was 31 ± 4 c.c. per minute; it increased abruptly to 97 ± 12 c.c. per minute when the circumflex coronary-pulmonary artery fistula was opened. Concurrently, the measured flow in the distal circumflex decreased from 28 + 3 to 21 ± 3 c.c. per minute, a relative steal of 26 per cent of the coronary perfusion in this artery. Phasic flow patterns are shown in Fig. 2. The flow pattern in the proximal coronary artery changed abruptly from the usual diastolic flow dominance to a continuous systolic and diastolic flow pattern very similar to the pattern recorded in the fistula itself. In the distal coronary artery there was a slight reduction in diastolic blood flow; mean flow signals obtained from another experiment (shown in Fig. 3) demonstrate the definite reduction in perfusion of the distal coronary bed. Left ventricular function. The hemodynamic data are summarized in Table I. Arterial pressure, left ventricular pressure and its derivative, stroke volume, heart rate, and cardiac output did not vary with the fistula open or closed. The coronary blood flow in the proximal coronary artery was
Volume 69 Number 2 February, 1975
Experimental
coronary
C-PAF CLOSED
PROXIMAL CORONARY (CC/min)
DISTAL CORONARY
(CC/min)
FISTULA (CC/min)
2 8 1
C-PAF OPEN
"^^^U4-uu^
ECG
artery fistula
^u*4-*
250 r 125 i-
250125 i0'
25C1250-J
EXP IR035
L SECONDS
Fig. 2. Recordings of hemodynamic parameters measured with the circumflex-pulmonary artery fistula (C-PAF) closed and after it had been opened for 2 minutes. ECG, Electrocardiogram.
MEAN DISTAL CORONARY FLOW (CC/min)
FISTULA OPEN
FISTULA CLOSED
250 r125 0
ECG
i?fi?N^5»f^i^^^ I
I
l _
i
EXP. 1032
i
i
i
i
i
i
i
i
i
i
i
i
SECONDS
Fig. 3. Mean distal coronary artery flow is shown as the fistula is opened for 5 seconds and then closed. ECG, Electrocardiogram. significantly increased (p < 0.001) and the flow in the distal coronary significantly decreased ( p < 0.02) with the shunt open. Discussion In these experiments, a coronary artery fistula was constructed to enter the pulmonary artery. It appears that the same method could be utilized to produce coronary arterial fistulas entering the right atrium, right ventricle, or left atrium. The
vein segments used ranged from 3 to 5 mm. in diameter and produced an average left-toright shunt of 1.1:1, a shunt comparable in magnitude to that measured in most asymptomatic patients." We presume that, with larger vein segments (as from the external jugular vein), larger shunts could be produced. It may be possible to examine the longterm effects of a coronary artery fistula with this model. The tendency for the proximal
The Journal of
2 82
Reitz, Harrison, Michaelis
Thoracic and Cardiovascular Surgery
coronary artery to dilate under conditions of increased flow,1 the chronic hemodynamic and morphologic effects of a fistula, and the natural history of such lesions are important areas to examine. It has been suggested that the greatly increased flow in the proximal coronary artery may contribute to accelerated atherogenesis.■'' If so, another use for the model would be to study this phenomenon. The presence of a steal phenomenon has been inferred from the characteristics of peripheral arteriovenous fistulas1^ and from angiography1' but has not been demonstrated directly. In these experiments with a small circumflex coronary-pulmonary artery fistula, there was a significant decrease (26 per cent) in distal coronary flow. This represents a relative reduction and not a reversal of flow, although flow reversal might be demonstrated with larger shunts. Under the conditions of this experiment, there was no demonstrable change in left ventricular function. This corresponds to the clinical situation in the majority of patients who are asymptomatic and who have normal catheterization findings except for the presence of a left-to-right shunt.' 7 - "' However, the effect of anesthesia, thoracotomy, and the acute nature of the experiments does not warrant the conclusion that the demonstrated decrease in distal coronary flow is without significant effect. REFERENCES 1 Sabbagh, A. H., Schocket, L. I., Griffin, T., Anderson, R. M., Goldberg, S., Fritz, J. M., and O'Hare, J.: Congenital Coronary Artery Fistula,
J.
794, 1973.
THORAC.
CARDIOVASC.
SURG.
66:
2 Parker, F. B., Neville, J. F., Johnson, L. W„ Scrivani, J. V., and Webb, W. R.: Congenital Coronary Artery Fistula From Supernumerary Coronary
Artery,
SURG. 65: 569,
J.
THORAC.
CARDIOVASC.
1973.
3 Jaffe, R. B., Glancy, D. L., Epstein, S. E., Brown, B. G., and Morrow, A. G.: Coronary Arterial-Right Heart Fistulae: Long Term Observations in Seven Patients, Circulation 47: 133, 1973. 4 Bravo, A. J., Glancy, D. L., Epstein, S. E., and Morrow, A. G.: Traumatic Coronary Arteriovenous Fistula, Am. J. Cardiol. 27: 673, 1971. 5 Edis, A. J., Schattenberg, T. T., Feldt, R. H., and Danielson, G. K.: Congenital Coronary Artery Fistula: Surgical Considerations and Results of Operation, Mayo Clin. Proc. 47: 567, 1972. 6 De Nef, J. J. E., Varghese, P. J„ and Losekoot, G.: Congenital Coronary Artery Fistula: Analysis of 17 Cases, Br. Heart J. 33: 857, 1971. 7 McNamara, J. J., and Gross, R. E.: Congenital Coronary Artery Fistula, Surgery 65: 59, 1969. 8 Nora, J. J., McNamara, D. G., Hallman, G. L., Sommervilk, R. J., and Cooley, D. A.: Medical and Surgical Management of Anomalous Origin of the Left Coronary Artery From the Pulmonary Artery, Pediatrics 42: 405, 1968. 9 Sabiston, D. C , Jr., Ross, R. S., Criley, J. M., Gaertner, R. A., Neill, C. A., and Taussig, H. B.: Surgical Management of Congenital Lesions of the Coronary Circulation, Ann. Surg. 157: 908. 1963. 10 Upshaw, C. B.: Congenital Coronary Arteriovenous Fistula: Report of a Case With an Analysis of Seventy-three Reported Cases, Am. Heart J. 63: 399, 1962. 11 Oldham, H. N., Ebert, P. A.. Young, W. G., and Sabiston, D. C , Jr.: Surgical Management of Congenital Coronary Artery Fistula, Ann. Thorac. Surg. 12: 503, 1971. 12 Rowe, G. G.: Inequalities of Myocardial Perfusion in Coronary Artery Disease ("Coronary Steal"), Circulation 42: 193, 1970.