Mechanism of the effect of coronary artery stenosis on coronary flow in the dog Kirk Lipscomb, M.D.* K. Lance Gould, M.D. Seattle, Wash.
Clinical studies have shown that the diseased coronary arterial system supplies normal coronary flow at rest; however, there is conflicting evidence as to whether this flow increases appropriately in response to hyperemic stjmuli.1-16 Although the study of this situation is difficult in man because of the inability to directly measure vessel flow, dog studies in which coronary flow was directly measured have shown that progressive stenosis initially limits the maximum hyperemic response, and only after the stenosis has progressed to the severity that this response is almost abolished, does it decrease resting flow.i7-lQ While this effect of stenosis on flow is known, the hemodynamic mechanism of this effect is unclear. Accordingly, the purpose of this study is to define this mechanism. To accomplish this, angiographic contrast media, an agent known to maximally but transiently vasodilate the coronary bed,lQ was injected into the variably stenotic dog coronary artery while the hemodynamic relationships of the stenosis and distal bed were studied separately and interdependently. Methods
Ten consecutive, 22 to 40 kilogram Black Labrador or German Shepherd dogs were studied. Each was given 45 mg. of morphine sulfate intramuscularly one hour prior to the procedure followed by intravenous sodium pentobarbital (20 mg. per kilogram) initially and as needed for anesthesia. Respiration was mainFrom the Cardiology Received
Department of Medicine, University Service, Veterans Administration for publication
April
of Washington, Hospital. Seattle.
8, 1974.
Reprint requests: Dr. Kirk Lipscomb, Veterans Administration tal, 4500 S. Lancaster Rd., Dallas, Texas 75216.
60
and
Hospi-
tained with a Harvard ventilator-y pump through a cuffed endotrachael tube, and blood PO, kept between 90 and 150 using supplemental oxygen, as necessary. The chest was entered through a left thoracotomy and the circumflex coronary artery was isolated. An electromagnetic flow-probe (Zepeda) with a lumen slightly smaller than the artery was placed just distal to the bifurcation from the left anterior descending artery. A variable constrictor was placed 0.5 cm. distal to the flow-probe. This constrictor consisted of a 3 mm. wide umbilical tape which passed around the artery and through an interposed length of stiff tubing to a micrometer such that the artery could be constricted in precise increments. Approximately 1.5 cm. distal to the constrictor and just proximal to the first major circumflex bifurcation, a 1 mm. outside diameter teflon end-hole catheter (Bardic 1966-T) was inserted pointing upstream; hereafter, this catheter is called the distal circumflex catheter. Any small branches between the flow-probe and distal circumflex catheter were ligated. A No. 8 French side-hole catheter was introduced through the right carotid artery and the end placed just above the aortic valve. In five dogs, a Sones coronary arteriography catheter was introduced through the left carotid artery. All measurements were recorded on an Electronics for Medicine DR-12 recorder at paper speeds varying from 25 to 100 mm. per second. Pressures were obtained through the aortic and distal circumflex catheters with a Kulite PSL 125-6 and a Statham P-23 Db pressure transducer, which were matched before, during, and after the procedure. Circumflex coronary flow
January, 1975, Vol. 89, No. 1, pp. 60-67
Effect of coronary
artery stenosis on , was recorded with no stenosis present and shows a slight gradient develop between the aorta and distal circumflex during hyperemia, while flow increases markedly. The lower tracing, (B), was recorded with a moderately severe stenosis, and in contrast to the unstenosed artery, shows the gradient increase markedly while flow increases relatively little. Initial control resting flow averaged 48 + 15 C.C. per minute, and after contrast injection increased 4.2 +- 0.4 times resting, a response 6 f 6 per cent greater than that following a lo-second occlusion. The heart rate was 151 f 2 beats per minute and the mean aortic pressure was 110 * 1 mm. Hg. Systemic changes after contrast injection were minimal averaging a decrease of 2 + 0.4 beats per minute in the heart rate and 5 _+ 0.5 mm. Hg in the mean aortic pressure. Verification of method. The flow-probe caused a 1.4 + 0.5 mm. Hg gradient at peak hyperemia and was disregarded in later calculations. The Sones catheter in the coronary orifice caused no significant gradient at peak hyperemia (0.5 f 0.6 mm.. Hg). Insertion of the catheter in the distal circumflex caused a 6 _t 5 per cent decline in peak flow after lo-second occlusion, a statistically insignificant amount. The stability of the preparation was documented by the peak flow
Results
Typical records of aortic pressure, distal circumflex pressure, and coronary flow recorded
62
January, 1975, Vol. 89, No. 1
Effect of coronary UFteFy
stenosison coronary flow
f
IO NORMALIZED
Fig. 4. Representative
gradient for one taken at maximum flow levels.
20
40
30
MEAN FLOW - TIMES
INITIAL
50
CONTROL
relationship between flow and pressure moderately severe stenosis. Points were hyperemia, at rest, and at intermediate
response decreasing only 13 t- 6 per cent from the beginning to the end of the experiment. Distal bed hemodynemics. The distal bed was defined as that part df the coronary vascular system distal to the distal circumflex catheter while the proximal coronary artery was defined as the artery proximal to the catheter. By measuring distal bed perfusion pressure through the distal circumflex catheter, the relationship of flow to distal bed pressure could be studied as this pressure was varied by changing the degree of stenosis. As shown in Fig. 2, resting coronary flow stayed at initial control level despite decreasing distal bed pressure until this pressure was dropped to approximately 60 mm. Hg. Below this pressure, flow was linearly related to distal bed pressure, a relationship which could be expressed as A distal bed pressure/A normalized flow, and termed distal bed resistance. Thus, in the resting state when distal bed pressure was reduced below 60 mm. Hg, distal bed resistance decreased to a minimal fixed value of 40 mm. Hg/normalized flow; but when pressure was increased above 60 mm. Hg, resistance was autoregulated such that resting flow was kept constant at initial control level. Since, at the minimal fixed resistance, the regression line correlating pressure and flow crossed the pressure axis at 16 mm. Hg, a point termed the critical closing pressure; the absolute relationship of pressure to flow was given by:
American Heart Journal
/ NORMALIZED
2 MEAN
FLOW -TIMES
4
3 INITIAL
5
CONTROL
Fq. 6. Regression lines relating pressure gradient to flow for 44 stenosea. Each line was &awn from 5 to 10 points, as shown in Fig. 4. The median correlation coefbient was 0.99. The degree of stenosis ranged from minimal (flat lines) to severe keep lines).
DBP = (DBR X F) + CCP
(1)
where: DBP = distal bed pressure; DBR = distal bed resistance (A distal bed pressure/A normalized flow); F = normalized flow; and CCP = critical closing pressure of distal bed. The relationship of flow to distal bed pressure at peak hyperemia was examined in similar fashion to that at rest, only the points were taken at the time of maximum response to contrast, one point for each stenosis. As shown in Fig. 3, the relationship was linear, indicating that at maximum vasodilation, distal bed resistance was constant, regardless of the severity of stenosis. This value was 20 mm. Hgjnormalized flow, a value less than the minimum resistance resulting from decreased pressure alone. However, the regression line representing this relationship also crossed the pressure axis at the critical closing pressure of 16 mm. Hg, such that Equation 1 could also be used to describe this relationship using the value of 20 mm. l&/normalized flow for distal bed resistance. Coronary steno& hemodynamiar. The isolated hemodynamics of the coronary stenosis were studied by plotting 5 to 10 points comparing the pressure gradient across the stenosis to flow through the stenosis as flow was transiently varied by contrast injection. An example of this relationship for one stenosis is shown in Fig. 4.
63
Lipscomb
and Gould
pressure gradient/A normalized flow); and FAI = flow axis intercept of the stenosis pressure gradient-flow regression line. Relationship of flow to stenosis resistance.
The theoretical relationship of coronary flow to coronary stenosis resistance was derived from the pressure-flow relationships of the distal bed coronary stenosis. By definition: AoP = DBP + PG where AoP = aortic pressure. Substituting mulas 1 and 2 into 3, and rearranging: F = AoP - CCP + (FAI x CSR) DBR + CSR 60 80 C%ONARY %ENOSIS RESISTANCE A mmHg/A Normalized Flow
100
Fig. 6. Effect of coronary stenosis resistance on resting flow (solid line, solid dots) and maximally hyperemic flow (dashed line, open circles). The lines represent the calculated values from Equation 5 while the points represent the observed values.
Forty-four stenoses of varying severity were studied in this manner and a composite of all their regression lines is illustrated by Fig. 5. All stenoses showed a highly linear relationship between pressure gradient and flow with a median correlation coefficient of 0.99. The slope of the regression line, A pressure gradient/A normalized flow, indicated the physiologic severity of the stenosis and was termed coronary stenosis resistance. An unexpected, but highly significant finding was that the regression lines relating pressure gradient to flow all intercepted the flow axis (zero pressure gradient) at a significantly positive flow. Although this intercept tended to be higher with low stenosis resistances and lower with high resistances, it averaged 0.65 f 0.03 times initial control flow for all stenoses in which a resting pressure gradient was present. Using this intercept and coronary stenosis resistance the relationship of pressure gradient to flow could be expressed as: PG = CSR X (F-FAD
(2)
where, PG = mean pressure gradient across stenosis; CSR = coronary stenosis resistance (A
64
(3) For-
(4)
Substituting the relatively constant aortic pressure, critical closing pressure, and flow axis intercept gave: F=
0.65 CSR f 94 DBR + CSR
(5)
In the resting state, flow was dependent on both stenosis and distal bed resistance at low and moderate values of stenosis resistance since distal bed resistance was autoregulated to keep flow at its control resting value. However, as stenosis resistance was increased, distal bed resistance decreased until it reached its minimum resting value of 46 mm. Hg/normalised flow. As stenosis resistance was increased beyond this point, flow was directly dependent on stenosis resistance and fell below its control resting value. The solid line in Fig. 6 illustrates that calculated resting flow remained at initial control level as stenosis resistance increased to the relatively severe value of 100, while the solid points indicate the observed values. Few data points are given at the highest resistance values because flow did not vary sufficiently to calculate a stenosis resistance. In the postcontrast, maximally hyperemic state, flow was dependent only on stenosis resistance since distal bed resistance was fixed. Therefore, any change in stenosis resistance resulted in a change in maximum flow. The dashed line in Fig. 6 indicates this predicted relationship from Formula 5, while the open circles indicate the observed values. Comparison of the two curves in Fig. 6 illus-
January, 1975, Vol. 89, No. 1
Effect of coronary
trat,es the capacity of the coronary arterial system to increase flow from the resting to the maximally hyperemic state and how this ratio decreased as coronary stenosis increased. In Equation 5, low, fixed levels of stenosis resistance were dominated by the higher values of distal bed resistance allowing changes in distal bed resistance to be reflected by relatively large changes in flow. In contrast, high, fixed levels of stenosis resistance dominated the equation such that changes in distal bed resistance were reflected by relatively small changes in flow. An additional factor which affected the change in flow from resting to hyperemia was that as stenosis resistance increased, resting distal bed resistance decreased to keep resting flow at control level. This decreased the amount of change in distal bed resistance when the resistance was changed from its resting to hyperemic state, thereby causing less change in flow. Discussion
Angiographic contrast media was used as the hyperemic stimulus in this study because of its maximum vasodilatory ability and rapid reversibility of effect. Contrast has been shown to cause maximum vasodilation since the peak flow after its injection is essentially the same as that following lo-second occlusion of the vessel,19 an observation confirmed in this study. Previous studies have shown that the peak flow rate after lo-second occlusion is the same as that after much longer occlusion and exceeds the flow rate with heavy exercise or excitement; and furthermore, is unaffected by vasodilator or beta-blocking agents, cardiac denervation, or simultaneous occlusion of the other coronary arteries.181 22-24 Although the dose of injected contrast was not precisely controlled due to the varying amount of myocardium supplied and some loss into the aorta, the similarity of flow responses suggested that the doses administered were on a plateau of the hypothetical dose-response curve. Clinical studies using isotopic methods to measure coronary flow after contrast injection have shown much lower flow responses to contrast,26 however, the methodology necessitated flow measurement much later after injection, at a time which our study would indicate was well after the peak response. The regression lines relating flow to distal bed
American
Heart Journul
artery stenosis on coronary
flow
crossed the pressure axis at approximately 16 mm. Hg, both in the resting state and after contrast. This pressure, where flow ceases, has long been recognized in skeletal muscle and termed “critical closing pressure.“27 More recently, Mosher and associates observed the same phenomenon in the heart using a coronary perfusion technique.28 With a given stenosis, the pressure gradient across the stenosis varied linearly with the flow through it, a finding predicted by Poiseulles equation.z7 However, an unexpected finding was that the regression line relating pressure gradient and flow did not intercept the flow axis at zero flow but, instead, intercepted it at a significantly positive flow. Although we are unable to explain this finding, support for its validity is gained from similar pressure gradient-flow curves in a postmortem study by Schultz, Hokanson, and Strandnesszg in which diseased femoral arteries were perfused. In all cases of significant stenosis, this flow axis intercept of the regression line was beyond the extent of collected data, therefore, we do not mean to imply that the pressure gradient necessarily decreases to zero before flow ceases. The position of the intercept is only used to identify the position of a regression line with a given slope. Decreased distal bed pressure has been shown to be an important cause of subendocardial ischemia.30-33 Since distal bed pressure is inversely related to the stenosis pressure gradient, which is in turn determined by the flow through the stenosis; Fig. 5 illustrates the critical effect that seemingly minor changes in flow through a severe stenosis have on the distal bed pressure. Through this mechanism, factors which increase coronary flow may lead to a “steal” phenomencn, i.e., increased subepicardial flow causes increased flow through the stenosis, increasing the pressure gradient and decreasing distal bed pressure, thereby resulting in decreased subendocardial flow. Conversely, factors which tend to decrease coronary flow may have an opposite effect, tending to improve subendocardial perfusion, pressure
Summary
The hemodynamic mechanism of the efIect of coronary artery stenosis on coronary flow was studied in the circumflex artery of 10 open-chest dogs by simultaneously measuring coronary flow,
65
Lipscomb
and Gould
aortic pressure, and coronary artery pressure distal to an adjustable constrictor; while the distal coronary bed was intermittently maximally vasodilated by intracoronary injections of angiographic contrast media (Hypaque-M, 75 per cent). For each stenosis, the pressure gradient across the stenosis varied directly with the flow through the stenosis (r = 0.99), the slope of the regression indicating the severity of the stenosis. An important observation was that this regression line did not intercept the flow axis at zero flow, but at a positive flow, meaning that for a given regression line slope the pressure gradient was much less than expected, At rest, distal bed resistance decreased as progressive stenosis lowered the distal bed pressure, maintaining flow at control level until the distal bed pressure dropped below 60 mm. Hg. However, at maximum hyperemia, distal bed resistance was at a fixed minimum value such that flow was directly proportional to distal bed pressure. Hence, progressive stenosis decreased the ratio of hyperemic to resting flow by: (1) causing the vasodilatory reserve to be used to maintain resting flow, decreasing that available for hyperemia, and (2) dropping the distal bed pressure relatively more for smaller increases in flow. This study provides a hemodynamic explanation for the known fact that progressive stenosis initially limits the maximum hyperemic flow, and only after this flow is decreased almost to resting level, does resting flow fall.
7.
8.
9.
10.
11.
12.
13. 14.
15.
16.
17.
The technical assistance of Miss Cynthia Calvert is gratefully acknowledged.
REFERENCES
1. Holmberg, S., Paulin, S., Prerovsky, I., and Varnauskas, E.: Coronary blood flow in man and its relation to the coronary arteriogram, Am. J. Cardiol. 19:486, 1967. 2. Schwartz, L., Froggatt, G., Cowey, H. D., Taylor, K., and March. J. E.: Measurement of left anterior descending coronary arterial blood flow: technique, methods of blood flow analysis, and correlation with angiography, Am. J. Cardiol. 32:679, 1973. 3. Knoebel, S. B., McHenry, P. L., Bonner, A. J., and Phillips, J. F.: Myocardial blood flow in coronary artery disease: effect of right atria1 pacing and nitroglycerin, Circulation 47:690, 1973. 4. Bing, R. J.. and Hellberg, K.: Coronary blood flow in relation to angina pectoris, Circulation 481146, 1972. 5. Yoshida, S., Ganx, W., Donoso, R., Marcus, H. S., and Swan, H. J. C.: Coronary hemodynamics during successive elevation of heart rate by pacing in subjects with angina pectoris, Circulation 44:1062, 1971. 6. Conti, R. C., Pitt, B., Gundel, W. D., Friesinger, G. C.,
66
and Ross, R. S.: Myocardial blood flow in pacing-induced angina, Circulation 42815, 1970. Benchimol, A., Desser, K. B., and Gartlan, J. L.: Effects of amyl nitrite on coronary arterial blood flow velocity in man, Am. J. Cardiol. 39:327,1972. Sullivan, J. M., and Gorlin, R.: Effect of I-epinephrine on the coronary circulation in human subjects with and without coronary artery disease, Circ. Res. 21:919, 1967. Cohen, L. S., Elliott, W. C., Klein, M. D., and Gorlin, R.: Coronary heart disease: clinical, cinearteriographic, and metabolic correlations, Am. J. Cardiol. 17:153, 1966. Holmberg, S., Serxysko, W., and Varnauskas, E.: Coronary circulation during heavy exercise in control subjects and patients with coronary heart disease, Acta. Med. Stand. 190~465, 1971. Parker, J. O., West, R. O., and Di Giorgi, S.: The effects of nitroglycerin on coronary blood flow and the hemody namic response to exercise in coronary artery disease, Am. J. Cardiol. 27:59, 1971. Knoebel, S. B., Elliott, W. C., McHenry, P. L., and Ross, E.: Myocardial blood flow in coronary artery disease: correlation with severity of disease and treadmill exercise response, Am. J. Cardiol. 27:51, 1971. Knoebel, S. B., McHenry, P. L., Phillips, J. F., and Paulette, F. J.: Coronary collateral circulation and myocardial blood flow reserve, Circulation 46:84, 1972. Ganz, W., Donoso, R., Marcus, H., and Swan, H. J. C.: Coronary hemodynamics and myocardial oxygen metabolism during oxygen breathing in patients with and without coronary artery disease, Circulation 45:763, 1972. Rowe, G. G., Thomsen, J. H., Stenlund, R. R., McKenna, D. H., Sialer, S., and Corliss, R. J.: A study of hemodynamics and coronary blood flow in man with coronary artery disease, Circulation 39:139, 1969. Forrester, J. S., Helfant, R. H., Pasternac, A., Amsterdam. E. A.. Most. A. S.. Kemu. H. G.. and Gorlin. R.: Atria1 pacing in coronary heartdisease: effect of hemodynamics, metabolism, and coronary circulation, Am. J. Cardiol. 27:237, 1971. Elliott, E. C., Jones, E. L., Bloor, C. M., Leon, A. S., and Gregg, D. E.: Day-to-&y changes in coronary hemodynamics secondary to constriction of circumflex branch of left coronary artery in conscious dogs, Circ. Res. 22:237,
1968.
18. Khouri, E. M., Gress, D. E., and Lowensohn, H. S.: Flow in the major branches of the left coronary artery during experimental coronary insufficiency in the unanesthetixed dog, Circ. Res. 23:99, 1968. 19. Gould, K. L., Lipscomb, K. M., and Hamilton, G. W.: Instantaneous flow response and regional distribution during coronary hyperemia as measures of coronary flow reserve, Am. J. Cardiol. 33:87, 1974. 20. Snedecor, G. W., and Cochran, W. G.: Statistical Methods, Ames, Iowa, 1967, Iowa State University Press. 21. Draper, N. R., and Smith, H.: Applied Regression Analysis, New York, 1966, John Wiley and Sons, pp. 71-72. 22. Olsson, R. A., and Gregg, D. E.: Myocardial reactive hyperemia in the unanesthetized dog, Am. J. Physiol. 208:224, 23.
24.
1965.
Gregg, D. E., Khouri, E. M., Donald, D. E., Lowensohn, H. S., and Pasyk, S.: Coronary circulation in the conscious dog with cardiac neural ablation, Circ. Res. 31:129, 1972. Pauly, T. J., and Bittar, N.: Myocardial reactive hyperemia responses in the dog after beta-receptor block with propranolol, Cardiovasc. Res. 5:440, 1971.
January,
1975, Vol. 89, No. 1
Effect of coronary
25.
26.
27.
28.
29.
Raff, W. K., Kosche, F., and Lochner, W.: Extravascular coronary resistance and its relation to microcirculation: influence of heart-rate, end-diastolic pressure, and maximal rate of rise of intraventricular pressure, Am. J. Cardiol. 29:598, 1972. Kloster, F. E., Friesen, W. G., Green, G. S., and Judkins, M. P.: Effects of coronary arteriography on myocardial blood flow, Circulation 46:438, 1972. Burton, A. C.: Physiology and Biophysics of the Circulation, Chicago, 1965, Year Book Medical Publishers, pp. 89-91. Mosher, B. A., Ross, J., Jr., McFate, P. A., and Shaw, R. F.: Control of coronary blood flow by an autoregulatory mechanism, Circ. Res. 14:250, 1964. Schultz, R. D., Hokanson, D. E., and Strandness, D. E., Jr.: Pressure-flow and stress-strain measurements of normal and diseased aorto-iliac segments, Surg. Gynecol. Obstet. 124:1267, 1967.
American
Heart Journal
30.
31.
32.
33.
artery
stenosis on. coronary
flow
Buckberg, G. D., Fixler, D. E., Archie, J. P., and Hoffman, J. L. E.: Experimental subendocardial ischemia in dogs with normal coronary arteries. Circ. Res. 30:67, 1972. Griggs, D. M., Jr., and Nakamura, Y.: Effect of coronary constriction on myocardial distribution of iodoantipyrine-1311, Am. J. Physiol. 215:1082, 1968. Moir, T. W., and De Bra, D. W.: Effect of left ventricular hypertension, ischemia, and vasoact.ive drugs on the myocardial distribution of coronary flow. Circ. Res. 21:65, 1967. Forman, R., Kirk, E. S., Downey, J. M., and Sonnenblick, E. H.: Nitroglycerin and heterogeneity of myocardial blood flow: reduced subendocardial blood flow and ventricular contractile force, J. Clin. Invest. 52:905, 1973.
67
Clinical studies have shown that the diseased coronary arterial system supplies normal coronary flow at rest; however, there is conflicting evidence as to whether this flow increases appropriately in response to hyperemic stjmuli.1-16 Although the study of this situation is difficult in man because of the inability to directly measure vessel flow, dog studies in which coronary flow was directly measured have shown that progressive stenosis initially limits the maximum hyperemic response, and only after the stenosis has progressed to the severity that this response is almost abolished, does it decrease resting flow.i7-lQ While this effect of stenosis on flow is known, the hemodynamic mechanism of this effect is unclear. Accordingly, the purpose of this study is to define this mechanism. To accomplish this, angiographic contrast media, an agent known to maximally but transiently vasodilate the coronary bed,lQ was injected into the variably stenotic dog coronary artery while the hemodynamic relationships of the stenosis and distal bed were studied separately and interdependently. Methods
Ten consecutive, 22 to 40 kilogram Black Labrador or German Shepherd dogs were studied. Each was given 45 mg. of morphine sulfate intramuscularly one hour prior to the procedure followed by intravenous sodium pentobarbital (20 mg. per kilogram) initially and as needed for anesthesia. Respiration was mainFrom the Cardiology Received
Department of Medicine, University Service, Veterans Administration for publication
April
of Washington, Hospital. Seattle.
8, 1974.
Reprint requests: Dr. Kirk Lipscomb, Veterans Administration tal, 4500 S. Lancaster Rd., Dallas, Texas 75216.
60
and
Hospi-
tained with a Harvard ventilator-y pump through a cuffed endotrachael tube, and blood PO, kept between 90 and 150 using supplemental oxygen, as necessary. The chest was entered through a left thoracotomy and the circumflex coronary artery was isolated. An electromagnetic flow-probe (Zepeda) with a lumen slightly smaller than the artery was placed just distal to the bifurcation from the left anterior descending artery. A variable constrictor was placed 0.5 cm. distal to the flow-probe. This constrictor consisted of a 3 mm. wide umbilical tape which passed around the artery and through an interposed length of stiff tubing to a micrometer such that the artery could be constricted in precise increments. Approximately 1.5 cm. distal to the constrictor and just proximal to the first major circumflex bifurcation, a 1 mm. outside diameter teflon end-hole catheter (Bardic 1966-T) was inserted pointing upstream; hereafter, this catheter is called the distal circumflex catheter. Any small branches between the flow-probe and distal circumflex catheter were ligated. A No. 8 French side-hole catheter was introduced through the right carotid artery and the end placed just above the aortic valve. In five dogs, a Sones coronary arteriography catheter was introduced through the left carotid artery. All measurements were recorded on an Electronics for Medicine DR-12 recorder at paper speeds varying from 25 to 100 mm. per second. Pressures were obtained through the aortic and distal circumflex catheters with a Kulite PSL 125-6 and a Statham P-23 Db pressure transducer, which were matched before, during, and after the procedure. Circumflex coronary flow
January, 1975, Vol. 89, No. 1, pp. 60-67
Effect of coronary
artery stenosis on , was recorded with no stenosis present and shows a slight gradient develop between the aorta and distal circumflex during hyperemia, while flow increases markedly. The lower tracing, (B), was recorded with a moderately severe stenosis, and in contrast to the unstenosed artery, shows the gradient increase markedly while flow increases relatively little. Initial control resting flow averaged 48 + 15 C.C. per minute, and after contrast injection increased 4.2 +- 0.4 times resting, a response 6 f 6 per cent greater than that following a lo-second occlusion. The heart rate was 151 f 2 beats per minute and the mean aortic pressure was 110 * 1 mm. Hg. Systemic changes after contrast injection were minimal averaging a decrease of 2 + 0.4 beats per minute in the heart rate and 5 _+ 0.5 mm. Hg in the mean aortic pressure. Verification of method. The flow-probe caused a 1.4 + 0.5 mm. Hg gradient at peak hyperemia and was disregarded in later calculations. The Sones catheter in the coronary orifice caused no significant gradient at peak hyperemia (0.5 f 0.6 mm.. Hg). Insertion of the catheter in the distal circumflex caused a 6 _t 5 per cent decline in peak flow after lo-second occlusion, a statistically insignificant amount. The stability of the preparation was documented by the peak flow
Results
Typical records of aortic pressure, distal circumflex pressure, and coronary flow recorded
62
January, 1975, Vol. 89, No. 1
Effect of coronary UFteFy
stenosison coronary flow
f
IO NORMALIZED
Fig. 4. Representative
gradient for one taken at maximum flow levels.
20
40
30
MEAN FLOW - TIMES
INITIAL
50
CONTROL
relationship between flow and pressure moderately severe stenosis. Points were hyperemia, at rest, and at intermediate
response decreasing only 13 t- 6 per cent from the beginning to the end of the experiment. Distal bed hemodynemics. The distal bed was defined as that part df the coronary vascular system distal to the distal circumflex catheter while the proximal coronary artery was defined as the artery proximal to the catheter. By measuring distal bed perfusion pressure through the distal circumflex catheter, the relationship of flow to distal bed pressure could be studied as this pressure was varied by changing the degree of stenosis. As shown in Fig. 2, resting coronary flow stayed at initial control level despite decreasing distal bed pressure until this pressure was dropped to approximately 60 mm. Hg. Below this pressure, flow was linearly related to distal bed pressure, a relationship which could be expressed as A distal bed pressure/A normalized flow, and termed distal bed resistance. Thus, in the resting state when distal bed pressure was reduced below 60 mm. Hg, distal bed resistance decreased to a minimal fixed value of 40 mm. Hg/normalized flow; but when pressure was increased above 60 mm. Hg, resistance was autoregulated such that resting flow was kept constant at initial control level. Since, at the minimal fixed resistance, the regression line correlating pressure and flow crossed the pressure axis at 16 mm. Hg, a point termed the critical closing pressure; the absolute relationship of pressure to flow was given by:
American Heart Journal
/ NORMALIZED
2 MEAN
FLOW -TIMES
4
3 INITIAL
5
CONTROL
Fq. 6. Regression lines relating pressure gradient to flow for 44 stenosea. Each line was &awn from 5 to 10 points, as shown in Fig. 4. The median correlation coefbient was 0.99. The degree of stenosis ranged from minimal (flat lines) to severe keep lines).
DBP = (DBR X F) + CCP
(1)
where: DBP = distal bed pressure; DBR = distal bed resistance (A distal bed pressure/A normalized flow); F = normalized flow; and CCP = critical closing pressure of distal bed. The relationship of flow to distal bed pressure at peak hyperemia was examined in similar fashion to that at rest, only the points were taken at the time of maximum response to contrast, one point for each stenosis. As shown in Fig. 3, the relationship was linear, indicating that at maximum vasodilation, distal bed resistance was constant, regardless of the severity of stenosis. This value was 20 mm. Hgjnormalized flow, a value less than the minimum resistance resulting from decreased pressure alone. However, the regression line representing this relationship also crossed the pressure axis at the critical closing pressure of 16 mm. Hg, such that Equation 1 could also be used to describe this relationship using the value of 20 mm. l&/normalized flow for distal bed resistance. Coronary steno& hemodynamiar. The isolated hemodynamics of the coronary stenosis were studied by plotting 5 to 10 points comparing the pressure gradient across the stenosis to flow through the stenosis as flow was transiently varied by contrast injection. An example of this relationship for one stenosis is shown in Fig. 4.
63
Lipscomb
and Gould
pressure gradient/A normalized flow); and FAI = flow axis intercept of the stenosis pressure gradient-flow regression line. Relationship of flow to stenosis resistance.
The theoretical relationship of coronary flow to coronary stenosis resistance was derived from the pressure-flow relationships of the distal bed coronary stenosis. By definition: AoP = DBP + PG where AoP = aortic pressure. Substituting mulas 1 and 2 into 3, and rearranging: F = AoP - CCP + (FAI x CSR) DBR + CSR 60 80 C%ONARY %ENOSIS RESISTANCE A mmHg/A Normalized Flow
100
Fig. 6. Effect of coronary stenosis resistance on resting flow (solid line, solid dots) and maximally hyperemic flow (dashed line, open circles). The lines represent the calculated values from Equation 5 while the points represent the observed values.
Forty-four stenoses of varying severity were studied in this manner and a composite of all their regression lines is illustrated by Fig. 5. All stenoses showed a highly linear relationship between pressure gradient and flow with a median correlation coefficient of 0.99. The slope of the regression line, A pressure gradient/A normalized flow, indicated the physiologic severity of the stenosis and was termed coronary stenosis resistance. An unexpected, but highly significant finding was that the regression lines relating pressure gradient to flow all intercepted the flow axis (zero pressure gradient) at a significantly positive flow. Although this intercept tended to be higher with low stenosis resistances and lower with high resistances, it averaged 0.65 f 0.03 times initial control flow for all stenoses in which a resting pressure gradient was present. Using this intercept and coronary stenosis resistance the relationship of pressure gradient to flow could be expressed as: PG = CSR X (F-FAD
(2)
where, PG = mean pressure gradient across stenosis; CSR = coronary stenosis resistance (A
64
(3) For-
(4)
Substituting the relatively constant aortic pressure, critical closing pressure, and flow axis intercept gave: F=
0.65 CSR f 94 DBR + CSR
(5)
In the resting state, flow was dependent on both stenosis and distal bed resistance at low and moderate values of stenosis resistance since distal bed resistance was autoregulated to keep flow at its control resting value. However, as stenosis resistance was increased, distal bed resistance decreased until it reached its minimum resting value of 46 mm. Hg/normalised flow. As stenosis resistance was increased beyond this point, flow was directly dependent on stenosis resistance and fell below its control resting value. The solid line in Fig. 6 illustrates that calculated resting flow remained at initial control level as stenosis resistance increased to the relatively severe value of 100, while the solid points indicate the observed values. Few data points are given at the highest resistance values because flow did not vary sufficiently to calculate a stenosis resistance. In the postcontrast, maximally hyperemic state, flow was dependent only on stenosis resistance since distal bed resistance was fixed. Therefore, any change in stenosis resistance resulted in a change in maximum flow. The dashed line in Fig. 6 indicates this predicted relationship from Formula 5, while the open circles indicate the observed values. Comparison of the two curves in Fig. 6 illus-
January, 1975, Vol. 89, No. 1
Effect of coronary
trat,es the capacity of the coronary arterial system to increase flow from the resting to the maximally hyperemic state and how this ratio decreased as coronary stenosis increased. In Equation 5, low, fixed levels of stenosis resistance were dominated by the higher values of distal bed resistance allowing changes in distal bed resistance to be reflected by relatively large changes in flow. In contrast, high, fixed levels of stenosis resistance dominated the equation such that changes in distal bed resistance were reflected by relatively small changes in flow. An additional factor which affected the change in flow from resting to hyperemia was that as stenosis resistance increased, resting distal bed resistance decreased to keep resting flow at control level. This decreased the amount of change in distal bed resistance when the resistance was changed from its resting to hyperemic state, thereby causing less change in flow. Discussion
Angiographic contrast media was used as the hyperemic stimulus in this study because of its maximum vasodilatory ability and rapid reversibility of effect. Contrast has been shown to cause maximum vasodilation since the peak flow after its injection is essentially the same as that following lo-second occlusion of the vessel,19 an observation confirmed in this study. Previous studies have shown that the peak flow rate after lo-second occlusion is the same as that after much longer occlusion and exceeds the flow rate with heavy exercise or excitement; and furthermore, is unaffected by vasodilator or beta-blocking agents, cardiac denervation, or simultaneous occlusion of the other coronary arteries.181 22-24 Although the dose of injected contrast was not precisely controlled due to the varying amount of myocardium supplied and some loss into the aorta, the similarity of flow responses suggested that the doses administered were on a plateau of the hypothetical dose-response curve. Clinical studies using isotopic methods to measure coronary flow after contrast injection have shown much lower flow responses to contrast,26 however, the methodology necessitated flow measurement much later after injection, at a time which our study would indicate was well after the peak response. The regression lines relating flow to distal bed
American
Heart Journul
artery stenosis on coronary
flow
crossed the pressure axis at approximately 16 mm. Hg, both in the resting state and after contrast. This pressure, where flow ceases, has long been recognized in skeletal muscle and termed “critical closing pressure.“27 More recently, Mosher and associates observed the same phenomenon in the heart using a coronary perfusion technique.28 With a given stenosis, the pressure gradient across the stenosis varied linearly with the flow through it, a finding predicted by Poiseulles equation.z7 However, an unexpected finding was that the regression line relating pressure gradient and flow did not intercept the flow axis at zero flow but, instead, intercepted it at a significantly positive flow. Although we are unable to explain this finding, support for its validity is gained from similar pressure gradient-flow curves in a postmortem study by Schultz, Hokanson, and Strandnesszg in which diseased femoral arteries were perfused. In all cases of significant stenosis, this flow axis intercept of the regression line was beyond the extent of collected data, therefore, we do not mean to imply that the pressure gradient necessarily decreases to zero before flow ceases. The position of the intercept is only used to identify the position of a regression line with a given slope. Decreased distal bed pressure has been shown to be an important cause of subendocardial ischemia.30-33 Since distal bed pressure is inversely related to the stenosis pressure gradient, which is in turn determined by the flow through the stenosis; Fig. 5 illustrates the critical effect that seemingly minor changes in flow through a severe stenosis have on the distal bed pressure. Through this mechanism, factors which increase coronary flow may lead to a “steal” phenomencn, i.e., increased subepicardial flow causes increased flow through the stenosis, increasing the pressure gradient and decreasing distal bed pressure, thereby resulting in decreased subendocardial flow. Conversely, factors which tend to decrease coronary flow may have an opposite effect, tending to improve subendocardial perfusion, pressure
Summary
The hemodynamic mechanism of the efIect of coronary artery stenosis on coronary flow was studied in the circumflex artery of 10 open-chest dogs by simultaneously measuring coronary flow,
65
Lipscomb
and Gould
aortic pressure, and coronary artery pressure distal to an adjustable constrictor; while the distal coronary bed was intermittently maximally vasodilated by intracoronary injections of angiographic contrast media (Hypaque-M, 75 per cent). For each stenosis, the pressure gradient across the stenosis varied directly with the flow through the stenosis (r = 0.99), the slope of the regression indicating the severity of the stenosis. An important observation was that this regression line did not intercept the flow axis at zero flow, but at a positive flow, meaning that for a given regression line slope the pressure gradient was much less than expected, At rest, distal bed resistance decreased as progressive stenosis lowered the distal bed pressure, maintaining flow at control level until the distal bed pressure dropped below 60 mm. Hg. However, at maximum hyperemia, distal bed resistance was at a fixed minimum value such that flow was directly proportional to distal bed pressure. Hence, progressive stenosis decreased the ratio of hyperemic to resting flow by: (1) causing the vasodilatory reserve to be used to maintain resting flow, decreasing that available for hyperemia, and (2) dropping the distal bed pressure relatively more for smaller increases in flow. This study provides a hemodynamic explanation for the known fact that progressive stenosis initially limits the maximum hyperemic flow, and only after this flow is decreased almost to resting level, does resting flow fall.
7.
8.
9.
10.
11.
12.
13. 14.
15.
16.
17.
The technical assistance of Miss Cynthia Calvert is gratefully acknowledged.
REFERENCES
1. Holmberg, S., Paulin, S., Prerovsky, I., and Varnauskas, E.: Coronary blood flow in man and its relation to the coronary arteriogram, Am. J. Cardiol. 19:486, 1967. 2. Schwartz, L., Froggatt, G., Cowey, H. D., Taylor, K., and March. J. E.: Measurement of left anterior descending coronary arterial blood flow: technique, methods of blood flow analysis, and correlation with angiography, Am. J. Cardiol. 32:679, 1973. 3. Knoebel, S. B., McHenry, P. L., Bonner, A. J., and Phillips, J. F.: Myocardial blood flow in coronary artery disease: effect of right atria1 pacing and nitroglycerin, Circulation 47:690, 1973. 4. Bing, R. J.. and Hellberg, K.: Coronary blood flow in relation to angina pectoris, Circulation 481146, 1972. 5. Yoshida, S., Ganx, W., Donoso, R., Marcus, H. S., and Swan, H. J. C.: Coronary hemodynamics during successive elevation of heart rate by pacing in subjects with angina pectoris, Circulation 44:1062, 1971. 6. Conti, R. C., Pitt, B., Gundel, W. D., Friesinger, G. C.,
66
and Ross, R. S.: Myocardial blood flow in pacing-induced angina, Circulation 42815, 1970. Benchimol, A., Desser, K. B., and Gartlan, J. L.: Effects of amyl nitrite on coronary arterial blood flow velocity in man, Am. J. Cardiol. 39:327,1972. Sullivan, J. M., and Gorlin, R.: Effect of I-epinephrine on the coronary circulation in human subjects with and without coronary artery disease, Circ. Res. 21:919, 1967. Cohen, L. S., Elliott, W. C., Klein, M. D., and Gorlin, R.: Coronary heart disease: clinical, cinearteriographic, and metabolic correlations, Am. J. Cardiol. 17:153, 1966. Holmberg, S., Serxysko, W., and Varnauskas, E.: Coronary circulation during heavy exercise in control subjects and patients with coronary heart disease, Acta. Med. Stand. 190~465, 1971. Parker, J. O., West, R. O., and Di Giorgi, S.: The effects of nitroglycerin on coronary blood flow and the hemody namic response to exercise in coronary artery disease, Am. J. Cardiol. 27:59, 1971. Knoebel, S. B., Elliott, W. C., McHenry, P. L., and Ross, E.: Myocardial blood flow in coronary artery disease: correlation with severity of disease and treadmill exercise response, Am. J. Cardiol. 27:51, 1971. Knoebel, S. B., McHenry, P. L., Phillips, J. F., and Paulette, F. J.: Coronary collateral circulation and myocardial blood flow reserve, Circulation 46:84, 1972. Ganz, W., Donoso, R., Marcus, H., and Swan, H. J. C.: Coronary hemodynamics and myocardial oxygen metabolism during oxygen breathing in patients with and without coronary artery disease, Circulation 45:763, 1972. Rowe, G. G., Thomsen, J. H., Stenlund, R. R., McKenna, D. H., Sialer, S., and Corliss, R. J.: A study of hemodynamics and coronary blood flow in man with coronary artery disease, Circulation 39:139, 1969. Forrester, J. S., Helfant, R. H., Pasternac, A., Amsterdam. E. A.. Most. A. S.. Kemu. H. G.. and Gorlin. R.: Atria1 pacing in coronary heartdisease: effect of hemodynamics, metabolism, and coronary circulation, Am. J. Cardiol. 27:237, 1971. Elliott, E. C., Jones, E. L., Bloor, C. M., Leon, A. S., and Gregg, D. E.: Day-to-&y changes in coronary hemodynamics secondary to constriction of circumflex branch of left coronary artery in conscious dogs, Circ. Res. 22:237,
1968.
18. Khouri, E. M., Gress, D. E., and Lowensohn, H. S.: Flow in the major branches of the left coronary artery during experimental coronary insufficiency in the unanesthetixed dog, Circ. Res. 23:99, 1968. 19. Gould, K. L., Lipscomb, K. M., and Hamilton, G. W.: Instantaneous flow response and regional distribution during coronary hyperemia as measures of coronary flow reserve, Am. J. Cardiol. 33:87, 1974. 20. Snedecor, G. W., and Cochran, W. G.: Statistical Methods, Ames, Iowa, 1967, Iowa State University Press. 21. Draper, N. R., and Smith, H.: Applied Regression Analysis, New York, 1966, John Wiley and Sons, pp. 71-72. 22. Olsson, R. A., and Gregg, D. E.: Myocardial reactive hyperemia in the unanesthetized dog, Am. J. Physiol. 208:224, 23.
24.
1965.
Gregg, D. E., Khouri, E. M., Donald, D. E., Lowensohn, H. S., and Pasyk, S.: Coronary circulation in the conscious dog with cardiac neural ablation, Circ. Res. 31:129, 1972. Pauly, T. J., and Bittar, N.: Myocardial reactive hyperemia responses in the dog after beta-receptor block with propranolol, Cardiovasc. Res. 5:440, 1971.
January,
1975, Vol. 89, No. 1
Effect of coronary
25.
26.
27.
28.
29.
Raff, W. K., Kosche, F., and Lochner, W.: Extravascular coronary resistance and its relation to microcirculation: influence of heart-rate, end-diastolic pressure, and maximal rate of rise of intraventricular pressure, Am. J. Cardiol. 29:598, 1972. Kloster, F. E., Friesen, W. G., Green, G. S., and Judkins, M. P.: Effects of coronary arteriography on myocardial blood flow, Circulation 46:438, 1972. Burton, A. C.: Physiology and Biophysics of the Circulation, Chicago, 1965, Year Book Medical Publishers, pp. 89-91. Mosher, B. A., Ross, J., Jr., McFate, P. A., and Shaw, R. F.: Control of coronary blood flow by an autoregulatory mechanism, Circ. Res. 14:250, 1964. Schultz, R. D., Hokanson, D. E., and Strandness, D. E., Jr.: Pressure-flow and stress-strain measurements of normal and diseased aorto-iliac segments, Surg. Gynecol. Obstet. 124:1267, 1967.
American
Heart Journal
30.
31.
32.
33.
artery
stenosis on. coronary
flow
Buckberg, G. D., Fixler, D. E., Archie, J. P., and Hoffman, J. L. E.: Experimental subendocardial ischemia in dogs with normal coronary arteries. Circ. Res. 30:67, 1972. Griggs, D. M., Jr., and Nakamura, Y.: Effect of coronary constriction on myocardial distribution of iodoantipyrine-1311, Am. J. Physiol. 215:1082, 1968. Moir, T. W., and De Bra, D. W.: Effect of left ventricular hypertension, ischemia, and vasoact.ive drugs on the myocardial distribution of coronary flow. Circ. Res. 21:65, 1967. Forman, R., Kirk, E. S., Downey, J. M., and Sonnenblick, E. H.: Nitroglycerin and heterogeneity of myocardial blood flow: reduced subendocardial blood flow and ventricular contractile force, J. Clin. Invest. 52:905, 1973.
67
