INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

SINCE Sabiston and Gregg (17) demonstrated that contraction of the heart impedes coronary arterial inflow, the characteristic phasic flow pattern of the coronary arterial system and the mechanisms behind this phenomenon have been studied extensively with reference to the susceptibility of the inner layer of the heart to myocardial ischemia (4, 7, 24). Although many investigators agree that the phasic flow pattern is related to the compression of intramural vessels, major controversy still persists as to the details of the mechanisms because no laboratory has reported the effects of cardiac contraction on intramural coronary diameters and flows, and this information is essential to an understanding of the influences of contraction on the phasic nature of perfusion.

The penetrating transmural arteries (PTAs) penetrate the heart wall from the outer to the inner layers, are surrounded by muscle fibers, and are affected by sagittally radiated left ventricular cavity pressure during systole or by strains along circumferential and longitudinal directions (16). Chilian and Marcus (3) and Kajiya et al. (10) reported systolic retrograde flow in the proximal portion of the septal artery. Ashikawa et al. (1) observed the configurational changes and flow pattern in small epicardial coronary vessels. They reported that vascular diameter in small subepicardial vessels is essentially unchanged throughout the cardiac cycle and that systolic forward flow is greater than in diastole. Judd and Levy (9) and Goto et al. (6) measured the diastolic and systolic blood volumes of intramural vessels in arrested hearts. These studies were static, and it is uncertain whether barium contracture is identical to normal contraction. Yada et al. (25) observed the configurational changes and flow pattern in small coronary vessels on epicardial and endocardial surfaces during a cardiac cycle, but not in intramural vessels. The intramyocardial pump model proposed by Hoffman and Spaan (7), Spaan et al. (19), and Flynn et al.(5) allows good organization of these observations. In this model, during systole blood is squeezed out of the intramural coronary vessels due to an abrupt increase in myocardial stiffness and intramyocardial pressure predominantly in the inner layer. Intramural vessels penetrating the heart (PTAs) are considered to be a common pathway for anterograde flow in diastole and retrograde flow in systole.

In the present study, we visualized configurational changes of the PTAs and the to-and-fro coronary blood flow (systolic elimination of contrast material followed by refilling in diastole) through PTAs by in situ microcoronary angiography using monochromatic synchrotron radiation (SR) and a high-definition (HD) video-camera system with a high-sensitivity image pick-up tube (8, 13, 14).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Surgical procedures and experimental protocol. In eight dogs (17.4 ± 1.2 kg body wt) anesthetized with morphine hydrochloride (3 mg/kg sc) and -chloralose (80 mg/kg iv), in situ coronary angiography was performed by using monochromatic SR with an energy of 33.3 keV and an HD video-camera system. A silicon tube bypass was set between the left subclavian artery and the left anterior descending artery after left thoracotomy and pericardiotomy. Coronary blood flow was monitored by an electromagnetic flowmeter prefixed in the arterial bypass. The left ventricular pressure was monitored with a catheter-tip manometer. In two of eight dogs, an additional bypass was placed between the distal great cardiac vein and right atrium for coronary venography. Two to three coronary arteriograms and/or two venograms were obtained with a 15-min interval between the injections; arteriograms were obtained at baseline in eight dogs, during intracoronary dobutamine (0.1 µg · kg¹ · min¹) in five dogs, and at coronary blood flow reduction to 40-50% of the baseline value in four dogs, and coronary venography was obtained at baseline in two dogs and during intracoronary administration of dobutamine in one dog. Two to three milliliters per second of iodine contrast material (iopamidol, Nihon Schering, Osaka, Japan, or iomeprol, Eisai, Tokyo, Japan) were injected into the bypass circuit via a three-way stopcock placed in the bypass for an injection period of 1.5-2.0 s while the dog was irradiated with 33.3 keV SR. The dogs were set nearly supine. The SR beam direction was set so as to pass through the left ventricular free wall from the posterobasal to the anteroapical direction. We also conducted fine adjustment of each dog's posture to obtain an optimal visual field; the target diagonal branch ran along the horizontal axis in the upper one-third of the visual field, and the monochromatic SR was nearly perpendicular to the virtual plane including the diagonal branch and its PTAs.

Microangiography using monochromatic synchrotron radiation. The coincidental SR at the beamline North-East-5 of the Accumulation Ring (ARNE5) or beamline 14-C from the Photon Factory (PF14) in the National Laboratory for High Energy Physics (Tsukuba, Japan) were monochromatized at 33.3 keV (just above the k-absorption edge of iodine) to optimize the detection of iodine and were magnified ×8-20 by means of an asymmetrically cut silicon crystal in front of the objects (Bragg reflection) (8). The angle of the coincidental beam direction and the lattice plane of the silicon crystal (Bragg angle: ) determine the energy of the diffracted monochromatic X-ray, and the angle of the lattice plane and the surface of the crystal () and the Bragg angle  determine the magnification ratio, as described previously (8, 13, 14). Monochromatic X-rays passing through the objects produced fluorescent images on the fluorescent screen. The visual field (scanning area on fluorescent screen) of the object was 2-3 cm by 2 cm and was not affected by the geometric relations among the light source, object, and detecting system, because monochromatic SR is a nearly parallel beam. These images were scanned by the HD video-camera system with a high-sensitivity image pick-up tube (avalanche-type video camera, either New Super Harp, Nippon Hoso Kyokai, or Harpicon, Hitachi, Tokyo, Japan) (12, 13, 21, 22) and then stored on the digital video system (HDD-1000, Sony, Tokyo, Japan) or a digital audio tape (C2594D, 2-8 GB, Hewlett-Packard, Palo Alto, CA) via a frame memory (12 bit/pixel) controlled by a work station (HP 9000 series 700, Hewlett-Packard). The image pick-up tube of the avalanche-type HD video camera contains an amorphous selenium-photoconductive target (12, 21) in which an electron-hole pair produced by an incident photon is accelerated by the application of a large electric field (avalanche phenomenon). The degree of multiplication indicated by an effective quantum ratio increases as the thickness of the selenium-photoconductive target, and the voltage of the electric field increases. This mechanism allows internal low-noise amplification up to >600 times that of the conventional video camera for the New Super Harp system and 30 times for the Harpicon system. A resolution bar chart (MICK type 14:2.0-20.0 line pairs/mm) study and a digital subtraction angiography phantom (type 76-700 Nuclear Associates, Carle Place, NY) study confirmed that this system can separate the adjacent lead line of 16 line pairs per millimeter (30 µm apart from each other) and can visualize vascular phantoms with a diameter of 500 µm and a minimum concentration of iodine (2.5 mg/ml) through a 7.5-cm-thick acrylic block. HD video-camera systems generally lose their sensitivity when their spatial resolution is improved, because the number of photons per unit area decreases as the spatial resolution increases. In this regard, the present HD video camera, which has an avalanche multiplication capacity, can be considered an ideal detecting system for obtaining precise vascular images without loss of sensitivity.

Assessment of configuration of penetrating transmural arteries and dynamic flow appearance. The configurational changes in PTAs and epicardial coronary arteries during a cardiac cycle were assessed on a computer (Power Macintosh 7100/80AV, Apple Computers, Cupertino, CA) using software in the public domain (NIH Image version 1.61) with some modifications after transfer from a digital memory source, as described in a previous paper (20). Before measurement, the angiographic images were processed using the linear interpolation method (15). This method allows reduction of effective pixel size to 9.8 × 9.8 µm by increasing the number of pixels in a 2 × 2-cm visual field to 2,048 × 2,048. The measured configuration indexes were the short-axial vascular diameter, as an index of vascular compression, and the longitudinal dimension of the proximal segment (a linear distance between the origin of the PTA and the first bifurcation point), reflecting vascular stretch. We selected a relatively late cardiac cycle in which PTAs were filled with contrast materials substantially for the configurational assessment. Vessel diameter is defined as the shortest distance between the two edges of a target vessel filled with contrast materials. First, we obtained the density profile at the measurement level. Local variation in the density profile was reduced by applying the running average method with a window value of 3 to 7 (15). The edges of the vessels were defined as the distance between the bilateral half-maximum points of the modified profile. In a preliminary study, the diameters of 37 vessels with a diameter range of 50-670 µm were independently measured by a medical student not involved in the experimental protocol and by a core member of the project. Regression analysis of the two sets of measurements yielded a regression coefficient of y = 1.02x  1.07 µm (r = 0.999, P < 0.001) and a very small standard deviation of the independent variables from the regression line (6.36 µm). Observation of the longitudinal vascular segment in a cardiac cycle allowed us to differentiate PTAs from the vessels on the surface, because PTAs (white arrowheads in Fig. 1) are stretched in systole, whereas surface vessels are shrunk (black arrowheads). The diameters of the PTA segments were measured at 62 sites with a mean depth of 3.2 ± 2.4 mm; sites were the midpoint of the origin of the PTA and the first bifurcation point, the midpoint of the first and second bifurcation points, or even lower levels. In 13 of the 62 segments, the diameter changes of the subendocardial segments at the level of 6.9 ± 0.9 mm from the epicardial surface (echocardiographically determined myocardial thickness of 8.0-9.0 mm) were compared with that of their proximal segments at the level of 1.5 ± 0.7 mm. Diameter changes of epicardial coronary arteries were determined at the midpoint of the origins of adjacent PTAs (n = 31). End diastole was defined as one or two frames (33-66 ms) before the time point at which the left ventricular pressure rose, and end systole was the end of the systolic plateau phase of the left ventricular pressure, based on simultaneously measured left ventricular pressure patterns. The longitudinal dimension of the proximal segment was defined as a linear distance between the origin of the PTA and the first bifurcation point.

View larger version (110K): [in this window] [in a new window]   Fig. 1.   Representative microangiograms of penetrating transmural arteries (PTA) arising from diagonal branches: diastole at baseline (A) and systole (B). Single, double, and triple white arrowheads indicate first-, second-, and third-order branches of large penetrating transmural arteries; 2 broad arrows indicate longitudinal dimension of proximal segment of a penetrating transmural artery; and black arrowheads indicate epicardial coronary artery.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

The present microangiographic system visualized a total of 189 PTAs (diameter range 60-650 µm) arising from the 9 diagonal branches as well as 7 distal portions of the left anterior descending artery in 8 dogs, more than 4 intramural coronary arteries (PTA) per 1-cm segment of each mother epicardial branch. We measured the diameter change of the vascular segments that had a diastolic value of 140-650 µm. The PTA, the branches of which are indicated by white arrowheads in Fig. 1, penetrated deep into the heart wall and probably fed deep myocardium. Its first-, second-, and third-order branches could be identified (single, double, and triple white arrowheads in Fig. 1). The detection rates of the first-, second-, and third-order branches of the 189 PTAs were 62, 17, and 3%, respectively. The two broad arrows in Fig. 1, A and B, indicate the longitudinal dimension of the primary segment defined by the distance between the PTA origin and the bifurcation point of the first-order branch. Comparison of angiograms at diastole (Fig. 1A) and systole (Fig. 1B) revealed that myocardial contraction compressed the PTAs (reducing the short-axial diameter) along the circumferential direction and stretched the vessels (increasing the longitudinal dimension) toward the ventricular cavity (hemodynamic values in Table 1). As shown in Fig. 2, the linear distance of the proximal PTA segments increased to 129 ± 25% of the diastolic value (n = 45, Fig. 2A), and in contrast, the short-axial diameters were reduced in systole to 87 ± 17% of the diastolic value (340 ± 110 µm, n = 62, Fig. 2B). As shown in Fig. 2C, the diameter of epicardial coronary vessels increased significantly in systole to 110 ± 8% of the end-diastolic value (from 700 ± 130 to 760 ± 150 µm, P < 0.01, paired t-test). The degree of vascular compression (%diameter change in systole) was more marked in the deep site than in the superficial site (60 ± 12 vs. 88 ± 12%, P < 0.001, paired t-test) in the 13 PTAs in which diameter could be measured at both the primary segment and the subendocardial segments (depths of measured sites: 1.5 ± 0.7 and 6.9 ± 0.9 mm, respectively). The percent change in diameter in systole correlated roughly with depth of measurement site from the epicardial surface (r = 0.51, P < 0.001, Fig. 3) but not with vessel size at end diastole (r = 0.11, Fig. 2B). Intracoronary administration of dobutamine (0.1 µg · kg¹ · min¹) also revealed significant changes in the degree of the reduction in the short-axial diameter (76 ± 12%, n = 19, P < 0.001, paired t-test) and the increase in the longitudinal dimension (138 ± 35%, n = 15, P < 0.001). In the groups of 19 and 13 PTAs in which contractile changes could be measured at both baseline and during intracoronary administration of dobutamine (see hemodynamic values in Table 1), the degrees of reduction in the short-axial diameters (90 ± 7 vs. 76 ± 12%, respectively, n = 19, P < 0.001, paired t-test) and the degree of increase in the longitudinal dimension (127 ± 37 vs. 143 ± 35%, respectively, P < 0.001, n = 13, paired t-test) were more marked during dobutamine administration than at baseline. Partial clamping of the proximal bypass circuit, causing a reduction of coronary blood flow to 40-50% of the baseline value (see hemodynamic values in Table 1), attenuated the relative changes in the short-axial diameters (96 ± 13%, n = 18, P < 0.05 with end-diastolic value, paired t-test) and longitudinal dimension (110 ± 20%, n = 21, not significant) in systole.

View larger version (17K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Fig. 2.   Relationships of degree of stretch of proximal PTA segments (A), relative diameter changes of PTA (B), and relative diameter changes of epicardial coronary segments (C) in systole to their end-diastolic values. As indicated by horizontal bars at level of 100%, proximal PTA segments increased longitudinally, PTA diameter decreased, and diameter of epicardial coronary segments increased. There was no significant correlation of either of the relative changes in systole to the end-diastolic value.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Correlation analysis of relative changes in PTA diameter to depth of vessels from epicardial surface. There is a significant negative correlation.

The to-and-fro appearance of coronary blood flow (slosh phenomenon) was more evident during partial clamping of the bypass circuit, causing a coronary blood flow reduction to 40% of the baseline value (Fig. 4), than at baseline or during dobutamine treatment. The systolic squeezing of blood in the PTA (Fig. 4, arrowhead) was extended even into the epicardial coronary artery (Fig. 4, arrow) beyond the origins of PTAs (compare Fig. 4, A and B). The difference between the contrast filling area of the PTAs in the two sequential end diastoles (Fig. 4, A and C, and D and F) indicates the effective forward flow toward the capillary beds (resistive flow) (2), and the difference between the end systole and the preceding end diastole (Fig. 4, A and B, and D and E) represents the amount of retrograde flow, which enhances the intravascular volume in the upper stream and does not contribute to exchange of gas and nutrients in the capillary beds (capacitive backflow) (2). The difference between the end systole and the next end diastole (Fig. 4, B and C, and E and F) is the sum of the capacitive and resistive flow. The degree of to-and-fro flow was not uniform among the PTAs; the resistive forward flow in the PTA at the most distal site (difference in contrast filling area of vessel as indicated by arrowhead in Fig. 4, D and F) is almost negligible and is apparently less than that in the PTA at the proximal site (indicated by arrowhead in Fig. 4, A-C). Systolic forward flow is noted only in the terminal epicardial segment (Fig. 4E, arrow), which did not have any intramural segments distally. Under baseline conditions and during intracoronary administration of dobutamine, the to-and-fro appearance of intramyocardial coronary blood flow was observed only in the deep portions of PTAs and did not extend beyond the proximal segments. In coronary venograms, systolic squeezing out of intramural blood was noted even in the baseline condition (Fig. 5). The entire intramyocardial venous segment was filled with contrast material in diastole (Fig. 5A). During systole the whole segment of PTAs became invisible (Fig. 5B), and in the subsequent diastole, refilling of these venous segments with the contrast was seen (Fig. 5C).

View larger version (102K): [in this window] [in a new window]   Fig. 4.   Angiograms showing to-and-fro appearance of intramural arterial flow during partial occlusion of coronary vessels. A-C: angiograms obtained during initial diastole, in subsequent systole, and then in next diastole, respectively, in early cardiac cycle. D-F: angiograms obtained during initial diastole, in subsequent systole, and then in next diastole, respectively, in late cardiac cycle. Arrowheads and arrows indicate visible terminal portion of PTAs and coronary arterial segments on epicardial surface. Wide arrows in A and B indicate increase of epicardial vessel diameter in systole.

View larger version (94K): [in this window] [in a new window]   Fig. 5.   Angiograms showing to-and-fro appearance of intramural venous flow under baseline condition. Angiograms were obtained during initial diastole (A), in subsequent systole (B), and then in next diastole (C). Arrowheads indicate intramural veins.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

Considerations of experimental model. The principal advantage of the present method is the visualization of intramural coronary vessels. Neither the floating-type microscope (1) nor needle-probe videomicroscope (25) can be applied to visualize intramural coronary vessels. In addition, the present method allows less invasive visualization of intramyocardial vessels and epicardial coronary artery. With the needle-probe videomicroscope, an inflated balloon is attached to the myocardial surface surrounding the target epicardial vessels. In addition, to observe endocardial arterioles, the needle has to be passed through the left ventricular cavity. With the floating-type microscope, the subepicardial muscle just below the target epicardial vessels must be punctured with a needle to fix the objective lens. One of the disadvantages of the present system is concomitant contrast material effects, which alter the vascular tonus and partially mask actual retrograde flow. In addition, the spatial resolution in the present system was poorer (9.8 µm in pixel size) than with the microscopic systems (1 and 5 µm, respectively). In the baseline and under dobutamine treatment, accompanying a high filling rate of contrast materials into the PTAs, the whole intramural arterial segments are filled with contrast materials within a single heart cycle. Under these conditions, the blood squeezing out of the distal PTAs already contains the contrast materials and makes it difficult to visualize the to-and-fro appearance. This might have affected the to-and-for appearance of PTAs under dobutamine administration, which was noted only in the deep (distal) PTA segments (see RESULTS).

New observations. In systole, the short-axial diameter of PTAs was reduced (compression along the circumference of the left ventricle) and their longitudinal dimension was increased (longitudinal stretching) toward the ventricular cavity (Figs. 1 and 2). The degree of compression correlated to the depth of the vessels with a predominance in the subendocardial segments (Fig. 3). These changes were enhanced by intracoronary administration of dobutamine. Partial clamping of the proximal site of the coronary arterial bypass almost abolished the systolic compression. The to-and-fro appearance of intramyocardial blood flow was visualized with an identification of capacitive backflow and resistive forward flow by the present angiographic system (Figs. 4 and 5). The relative magnitudes of resistive and capacitive flow predict the degree of to-and-fro flow in each PTA. Under partial occlusion of the proximal coronary artery, the relative magnitudes of these flows were not uniform among the PTAs (compare the vessel indicated by arrowheads in Fig. 4, A-C, and that in D-F). On coronary venography, marked squeezing out of intramural blood was noted even in the baseline condition (Fig. 5).

Comparisons of present results with previous reports. The present results could be interpreted as direct evidence supporting the classic hypothesis of Scaramucci (18) in 1695 that "the myocardial vessels are squeezed by the contraction of the muscle fibers around them" and the newest hypothesis of the systolic-diastolic interaction model in coronary circulation advocated by Spaan et al. (19), Hoffman and Spaan (7), and Flynn et al. (5). The longitudinal stretching of the vessels noted in the present study (Fig. 2A) allows the extravascular pressure to be transmitted to the intramural vessels efficiently. Flynn et al. (5) speculated that the blood ejected from the endocardium flows into epicardial small vessels during systole and that aortic pressure acts as a back pressure to the reverse flow into extramural vessels under baseline conditions. This hypothesis was confirmed by the direct visualization of retrograde flow from PTAs into epicardial vessels (Fig. 4) and the vascular diameter changes in intramural and epicardial vessels (Fig. 2). However, the systolic forward flow in the epicardial vessels reported by Ashikawa et al. (1) and by Yada et al. (25) was noted only in the terminal segment without intramural branches during partial clamping of the coronary bypass circuit. The more marked emptying of the intramural veins during systole (Fig. 5) is compatible with the description by Wiggers (24) that "the volume of blood that can enter intramural vessels during diastole must depend to some extent on the degree to which they are emptied during preceding systole." The lower back pressure for the anterograde intramural venous flow (right atrial pressure) than that for the retrograde PTA flow (aortic pressure) appears to be a major reason for this phenomenon. The degree of diameter change in the subepicardial PTA segment (88% of diastolic value, on average), measured at a depth of 1.5 ± 0.7 mm, was quite different from changes in small epicardial vessels (negligible changes throughout cardiac cycle) reported by Ashikawa et al. (1) using a floating-type microscope. Changes in the subendocardial segments (60% of diastolic value), measured at a depth of 6.9 ± 0.9 mm, was more marked than change seen in the endocardial arteriole by Yada et al. (25) employing a needle-probe videomicroscope with a charge-coupled device camera (20% reduction). Ashikawa et al. (1) reported that vascular diameter in small subepicardial vessels, i.e., those with a diameter 20 µm, is essentially unchanged throughout the cardiac cycle, and systolic forward flow is greater than in diastole. Yada et al. (25) also reported similar diameter changes in the epicardial arteriole. However, in the present study, the diameters of proximal segments of PTAs were significantly reduced. Judd and Levy (9) reported that the vascular volume in the diastolic arrested heart was larger than in the systolic arrested heart, even in the subepicardium. This discrepancy and the more marked diameter reduction in the subendocardial vessels than in the report by Yada et al. (25) might be related to the anatomic differences in the observed vessels. We visualized subepicardial segments of PTAs that were running through the heart wall toward the left ventricular cavity and, therefore, that were being compressed nearly perpendicularly by the heart muscle along their short axes. In contrast, Yada et al. (25) and Ashikawa et al. (1) observed the vessels apparently running parallel to the epicardial or endocardial surface. The difference in the size of the observed vessels might account, to some extent, for the difference in the degree of the compression between ours and those observed by Ashikawa et al. (1). The diameters of the vessels we observed (340 ± 10 µm) were much larger than those (20 µm) studied by Ashikawa et al. (1). However, the difference between our data and those of Yada et al. (25) was modest (169 ± 12 µm). Yada et al. (25) reported significant size dependency in the degree of reduction in vascular diameter by measuring the vessels ranging mainly from 50 to 250 µm. There was an obvious difference in the degree of reduction in the vessel diameter of <100 µm and that of >100 µm. In contrast, our measurements did not include so-called resistance vessels with diameters of <100 µm, which were characterized by fewer diameter changes than the larger segments. An alternative explanation is related to the technical inherency of the methods. The puncture of subepicardial muscle (1) or balloon attachment on the myocardial surface (25) might have reduced the extravascular compression to small coronary vessels.