Physiologic Characteristics of Coronary Artery Bypass Grafts Takeo Tedoriya, MD, Michio Kawasuji, MD, Keishi Ueyama, MD, Naoki Sakakibara, MD, Hirofumi Takemura, MD, and Yoh Watanabe, MD Department of Surgery (I), Kanazawa University School of Medicine, Kanazawa, Japan To investigate the hemodynamic characteristics of arterial grafts for coronary artery bypass grafting, we measured phasic pressure and flow patterns in three types of grafts in a canine model (n = 18). A graft from the ascending aorta (AAG), a graft from the descending aorta at the first lumbar level (DAG), analogous to a right gastroepiploic artery, and an internal thoracic artery (ITA) were anastomosed to each other. The composite graft was anastomosed to the left anterior descending coronary artery, and then the left anterior descending coronary artery was ligated. Before grafting, the AAG showed high sustained diastolic pressure, but the ITA and DAG showed rapid fall of diastolic pressures. Mean diastolic pressures were 83 f 2, 72 f 3, and 57 + 2 mm Hg in the AAG, ITA, and DAG (p < 0.05). Free flow in the AAG was markedly greater than in the ITA or the DAG. After grafting to the left anterior descending coronary artery, no changes were observed in diastolic pressures compared with the pregrafting values. Total blood flows were 72 2 6, 80 + 7, 57 f 7, and 44 f 6 ml/min in the left anterior descending coronary artery, AAG, ITA, and DAG, respectively. There were no differences in systolic graft flow between the three types of grafts. Diastolic blood flow in the ITA (29 f 4 ml/min) and DAG (18 f 3 ml/min) was smaller than in the AAG (48 f 4 ml/min) (p < 0.01). Regression equations between mean diastolic pressure (X) and diastolic graft flow (Y) were Y = 0.46X + 8.6 (T = 0.79; p < 0.01) in the AAG, Y = 1.3X 68 (T = 0.83; p < 0.001) in the ITA, and Y = 1.2X 54 (T = 0.91; p < 0.0001) in the DAG, respectively. These results indicate that arterial grafts have disadvantages over ascending aortaoriginated grafts in the ability to supply blood to the diastolicdominant coronary circulation. ( 1993;56:9516) he internal thoracic artery (ITA) is recognized as T having a superior rate of longterm patency and is the currently preferred conduit for coronary artery bypass grafting [l]. The flow rate of the ITA graft is adequate to relieve angina in most patients. However, there have been concerns about the possible inadequacy of flow through arterial grafts [251. Clinical conditions suggesting inadequate perfusion by the ITA graft have been For editorial comment, see page 809. recognized during and immediately after operation [6, 71. Jones and associates [6] and von Segesser and colleagues [7] separately reported that patients who received one or more ITA grafts showed signs of left ventricular failure, suggesting ITA hypoperfusion. They stated that the most important contributing factor of ITA hypoperfusion appeared to be a disproportion between ITA flow and myocardial demand. We found a slightly greater potential for inadequate flow in patients with ITA grafts, after comparing flow capacities of ITA and saphenous vein grafts under exercise conditions [5]. The right gastroepiploic artery (GEA) recently has been used as the third Accepted for publication Jan 27, 1993. Address reprint requests to Dr Kawasuji, Department of Surgery (I), Kanazawa University School of Medicine, Takaramachi 131, Kanazawa 920, Japan. arterial graft for coronary artery bypass operation [8, 91. It is probable that arterial grafts, when transferred to the heart, will not function like the coronary artery in some conditions. The purpose of the present study was to examine hemodynamic properties of ascending aortaoriginated grafts, ITA grafts, and GEA grafts by measuring their pressures and flow patterns to investigate flow capacities of arterial grafts and mechanisms of hypoperfusion. Material and Methods Eighteen adult mongrel dogs weighing 22 to 30 kg were studied. The animals were anesthetized with intravenous administration of ketamine hydrochloride (20 mg/kg) and were mechanically ventilated with a volume respirator. The left ITA was dissected from the thoracic wall through a left parasternal thoracotomy. All side branches of the ITA were ligated from the subclavian artery to the bifurcation of the musculophrenic artery. The ITA was sprayed and flushed with diluted papaverine chloride (0.4 mg/20 ml) to prevent spasm. The mean diameter of the ITAs was 2.5 mm. A femoral artery was excised in 2cm lengths and prepared for bypass grafts by flushing with diluted papaverine chloride. The mean diameter of the femoral artery grafts was 3 mm. A bubble oxygenator and a roller pump were equipped for cardiopulmonary bypass, and blood flow was diverted 0 1993 by The Society of Thoracic Surgeons 00034975/93/$6.00
952 TEDORIYA ET AL 1993;569514 with an aortic cannula for arterial inflow and superior and inferior venae cavae cannulas for venous return. After aortic crossclamping, cardiac arrest was induced with administration of cold crystalloid potassium cardioplegic solution. The femoral artery graft was anastomosed to the left anterior descending coronary artery (LAD) with a continuous 70 polypropylene suture. The mean diameter of the LADS was 2 mm at the anastomosis. After release of the aortic clamp, a polytetrafluoroethylene (PTFE) graft (3 mm in diameter) was anastomosed to the ascending aorta. In most dogs, the right GEA was about 1 mm in diameter and was not adequate for grafting. Another PTFE graft was anastomosed to the descending aorta at the level of the first lumbar vertebra. In this bypass model, a graft from the descending aorta (DAG) was anastomosed in lieu of an in situ GEA graft. A Yshaped graft was constructed from the two PTFE grafts of sufficient lengths to reach the LAD. The ITA was anastomosed to this Yshaped graft. Dogs were weaned from cardiopulmonary bypass after rewarming and achieved stable hemodynamics. Before the composite grafts were connected to the LAD, distal pressure in the three grafts was measured simultaneously with the left ventricular pressure. Free flows from the three grafts were measured by a timed volumetric collection. The composite graft was connected to the proximal portion of the femoral artery graft, which was already anastomosed to the LAD. The blood flow in the LAD was measured during clamping of the composite graft. After ligation of the LAD at the proximal portion of the anastomosis, graft flow was measured in the LAD immediately distal to the anastomosis. Pressure in the three types of grafts was measured at the end of the composite graft (Fig 1). The use of PTFE reduced the aortic clamp time and invasion for the experimental animals. Because PTFE was too hard to be anastomosed with the LAD, the femoral artery graft was interposed between the LAD and the composite graft. This experimental model allowed evaluation of the blood flow in the three types of grafts for the same coronary bed. After the operations, the configurations of the composite grafts were investigated to ensure that there was no stenosis and kinking. Hemodynamic properties in this model could not be altered in comparison with those of the same grafts being directly anastomosed to a coronary artery. Each measurement was performed over 5 minutes, in the following sequence: ascending aorta graft (AAG), ITA, and DAG. After these measurements were made, the distal pressure in the three types of grafts was remeasured with and without clamping of the end of the composite graft, so that the differences in pressure waves between the pregrafting and postgrafting states could be investigated. All of the animals were studied in the same order. Arterial and left ventricular pressures were measured by transducertipped pressure monitoring catheters [lo] (Camino, San Diego, CA). This fiberoptic transducer system operates via a light transmitting/receiving system that responds to the movement of a mirror diaphragm. pres mon Fig 1. Experimental model of coronary artery bypass grafting and the method for measuring pressure and blood pow. (AAG = ascending aorta graft; DAG = descending aorta graft; ITA = internal thoracic artery; LAD = left anterior descending coronary artery.) The light, which is sent and received, is analyzed by a microcomputer and converted into signals directly related to applied pressure. This fiberoptic system eliminates the artifacts inherent in a fluidfilled hemodynamic monitoring system. This system has enhanced response characteristics and improves the quality of the measured pressure waveforms. Blood flow was measured by a transittime ultrasonic blood flow meter [ll](model T101; Transonic System Inc, Ithaca, NY). The transducer emits a planewave ultrasonic burst, which intersects the full area of the vessel twice on a reflected pathway by the acoustic reflector. The transit time of the ultrasonic wave is slightly modified by the flow in the vessels. The direction of the transittime measurement alternates between the two transducers. "Upstream" and "downstream" phase readings are used to calculate to flow volume. In comparison with an electromagnetic flow meter, the prominent features of this system are direct volume flow metering, nonconstrictive fit around a vessel, and zero baseline stability. The transducer probe was fixed around the LAD, and the ultrasonic field of the probe was filled with jelly for stable ultrasound transmission, so that clear flow waves were displayed even for beating hearts. All recordings were made on a Mingograph 82 (SiemensElema, Stockholm, Sweden) multichannel recorder. The delay between the onset of systole, as determined by the peak of the R wave on the electrocardiogram and the peak systolic spike of the pulse pressure tracing, was measured in each graft. The areas under the systolic and diastolic pressure wave forms were measured by a planimeter, using the peak of the R wave and the end of T wave on the electrocardiogram as the references for systole. Mean
1993;569514 TEDORIYA ET AL 953 ECG ECG ECG mmhg mmhg mmhg 120 100 80 60 40 20 0 120, 120, l0omm Fig 2. Tip pressure waves of three types of grafts and left ventricular pressure. (AAG = ascending aorta graft; DAG = descending aorta graft; ECG = electrocardiogram; ITA = internal thoracic artery; LV = left ventricle.) n CT) 120 I 100 E v 80 al 2 60 v) v) 40 P = 20 0 0 S O ns systolic diastolic mean ns ** z nn mean mean systolic diastolic AAG ITA 0 DAG Fig 3. Tip pressures in three types of grafts. Values represent the mean t the standard error of the mean. (AAG = ascending aorta graft; DAG = descending aorta graft; ITA = internal thoracic artery; ns = not significant; * p < 0.05; ** p < 0.01.) systolic and diastolic pressures were calculated from systolic and diastolic areas divided by time. Dogs received humane care in compliance with the Principles of Laboratory Animal Care (National Society for Medical Research) and the Guide for the Care and Use of Laboratory Animals (National Academy of Sciences, NIH publication 8523, revised 1985). Mean values were calculated from pressures and flows measured during five consecutive cardiac cycles. Values are expressed as the mean f the standard error of the mean. Statistical analysis was performed with Student s t test and analysis of variance to detect significant ( p < 0.05) differences between measured variables. Results There were no changes in systemic blood pressure, heart rate, ST segments on the electrocardiogram, or ventricular wall motion during the measurement of the three types of grafts. There were no manifestations of myocardial ischemia. Free flows were 352 * 12, 93 * 5, and 80 2 6 mwmin from the AAG, ITA, and DAG, respectively. Free flow from the AAG was significantly greater than from the ITA and DAG (p < 0.01). Before grafting, tip pressures of the three types of grafts showed unique patterns. The AAG showed high sustained diastolic pressure that fell gradually. The pressures in the ITA and DAG had narrow systolic contours and fell rapidly. The pressure wave in the ITA had a notch between the systolic and diastolic contours (Fig 2). Systolic pressures were 109 f 2, 108 f 2, and 110 * 3 mm Hg and diastolic pressures were 73 f 2, 60 * 3, and 50 f 2 mm Hg in the AAG, ITA, and DAG, respectively. There were no differences in systolic pressures between the three types of grafts. The diastolic pressure was significantly lower in the ITA than in the AAG, and was significantly lower in the DAG than in the AAG and ITA. Mean pressures were 85 * 2, 77 * 2, and 65 f 3 mm Hg in the AAG, ITA, and DAG, respectively. The mean pressure was significantly lower in the DAG than in the AAG and ITA. Mean systolic pressures were 97? 2, 95 f 4, and 94 f 3 mm Hg and mean diastolic pressures were 83 f 2, 72 f 3, and 57 * 2 mm Hg in the AAG, ITA, and DAG, respectively. There were no differences in mean systolic pressures between the three types of grafts. Mean diastolic pressures in the ITA and DAG were significantly lower than in the AAG, and the mean diastolic pressure in the DAG was lower than in the ITA (Fig 3). Before grafting, the delay in pulse pressures were 82 * 6,130 f 4, and 177 f 5 milliseconds in the AAG, ITA, and DAG, respectively. All pulse delays were significantly (p < 0.01) different from each other (see Fig 2). After grafting to the LAD, the onset of systolic pressures in the three types of grafts was simultaneous with the left ventricular pressure due to the direct influence of the contraction of the left ventricle (Fig 4). There were no changes in the delay of pulse pressure before or after grafting. Peak systolic pressures after grafting were 95 f 3, 96 f 2, and 98 2 3 mm Hg in the AAG, ITA, and DAG, respectively before grafting +after grafting 1 sec Fig 4. The changes in intraluminal pressure in an internal thoracic artery (ITA) graft. Pressure waves were recorded continuously. Internal thoracic artery pressures before and after grafting showed intraluminal pressures of the closed and opened graft. (ECG = electrocardiogram; LV = left ventricular pressure.)
954 TEDORIYA ET AL 1993;56:9514 0 LAD @ AAG @ ITA @ DAG 1 sec Fig 5. Blood flows in the native LAD and in the three types of grafts. (AAG = ascending aorta graft; DAG = descending aorta graft; ECG = electrocardiogram; ITA = internal thoracic artey; LAD = left anterior descending corona y artery.) (not significant), and all were equalized with the left ventricular systolic pressure. The diastolic pressures in the three types of grafts were not different from their pregrafting values. The blood flow in the LAD had systolic and diastolic components. The blood flow wave in the AAG had a prominent diastolic component, similar to the native LAD. However, the blood flow in the ITA and the DAG had decreased diastolic components (Fig 5). Total blood flows were 72 6, 80? 7, 57 * 7, and 44? 6 ml/min in the LAD, AAG, ITA, and DAG, respectively. Total blood flow in the ITA was lower than in the A,4G, but this difference did not reach statistical significance. Total blood flow in the DAG was significantly lower than in the LAD and AAG. Systolic flows were 30 2 3, 32 2 5, 28 * 3, and 26 a 2 5 ml/min in the LAD, AAG, ITA, and DAG, respectively (not significant). Diastolic flows were 42 * 4, 48 * 4,29? 4, and 18 * 3 ml/min in the LAD, AAG, ITA, and DAG, respectively. Diastolic flows were significantly lower in the ITA and DAG than in the AAC (Fig 6). The relationship between tip pressures before grafting and diastolic graft flows was studied in the three types of $ I ** LAD AAG ** 0 DAG T I $* r1. // total flow systolic flow diastolic flow Fig 6. Comparison of total, systolic, and diastolic blood flows in the native left anterior descending coronary artery (LAD) i7nd in the three types of grafts. Values represent the mean f the standard error of the mean. (AAG = ascending aorta graft; DAG = descending aorta graft; ITA = internal thoracic artery; ns = not significant; * p < 0.05; ** p < 0.01.) 55 1 n 50. c E 45 1 E 40 v 4 + 25! ITAA / cn 20 0 Y =1.3X 68 r=0.83 p<o.ool o 15 Y v) U 7 Y =0.46X+8.6 r=0.79 p<o.ol Y =1.2X54 r=0.91 p<o.oool 5 50 60 70 80 90 mean diastolic pressure (rnrnhg) Fig 7. Correlation between diastolic blood flows and mean diastolic pressures of the three types of grafts. (AAG = ascending aorta graft; DAG = descending aorta graft; ITA = internal thoracic artery.) grafts. Regression equations between mean diastolic pressure (X) and diastolic graft flow (Y) were Y = 0.46X + 8.6 (r = 0.79; p < 0.01) in the AAG, Y = 1.3X 68 (Y = 0.83; p < 0.001) in the ITA, and Y = 1.2X 54 (r = 0.91; p < 0.0001) in the DAG, respectively (Fig 7). Comment The present study investigated physiologic characteristics of the three types of grafts after coronary artery bypass grafting. The blood flow in the AAG showed a twopeaked wave form, which was diastolicdominant, as in the native LAD flow. Blood flows in the ITA and DAG showed a remarkably decreased diastolic component compared with the native LAD flow. There were no differences in the systolic graft flow between each type of graft. The diastolic flow in the ITA and DAG was significantly lower than in the AAG. Furthermore, the mean diastolic pressures were highly correlated with the diastolic graft flows after anastomosis of the three types of grafts. Flow through a coronary bypass graft depends on many factors, including systemic hemodynamics [ 121, anatomic factors (length and diameter) and vasoactivity of the graft [13], coronary vascular resistance, and severity of stenosis in the native coronary artery. During constant distal coronary vascular resistance, the pressure gradient across the anastomosis will determine the coronary bypass inflow. Because coronary blood flow distal to the grafts occurs mainly in the diastolic phase [14], the diastolic pressure gradient across the anastomosis is the important driving pressure for coronary blood flow. The present study showed the physiologic characteristics of the grafts under stable systemic hemodynamics, constant vasoactivity of the graft, constant coronary vascular resistance, and total occlusion of the native coronary artery. In the present study, PTFE was substituted for GEA and an
1993;56:9516 TEDORIYA ET AL 955 AAG. The characteristics of the distal pressure waves in the ascending aortaoriginated saphenous vein graft, ITA, and GEA in our clinical study [15] were analogous to those in the AAG, ITA, and DAG, respectively, in this study. Differences in compliance and caliber between PTFE grafts in this model and in situ vein grafts or GEA could not have influenced the hemodynamic results in this study. Pressure wave forms of ITA and DAG showed decreased diastolic pressure. Mean diastolic pressures decreased in sequence: AAG, ITA, and DAG. The anatomic distances from the heart should contribute considerably to the observed pressure phenomena. Consequently, the diastolic graft flows decreased in the same sequence. The arterial pulse in the ITA and DAG was delayed, compared with the AAG, and this delay could have increased blood flow delivery during diastole. However, the systolic spike of pulse pressure in the ITA and DAG had narrow contours and did not enter during diastole. Therefore, pulse delay was not a factor in determining graft flow. Wakabayashi and associates [16] reported a significant correlation between diastolic graft pressure and graft flow in their experimental study. In a clinical study, GonzalesSantos and coworkers [17] also reported a significant correlation between diastolic graft pressures and ITA flow. Singh and associates [2] demonstrated that left ventricular enddiastolic pressure influenced graft flow, especially in the ITA. These results suggested that diastolic pressure should be considered the driving pressure for coronary circulation. Simon and colleagues [18] reported that after cardiopulmonary bypass, the left ventricle showed significant deterioration in ITArevascularized myocardial segments, more so than in saphenous veinrevascularized segments. Jones and coworkers [6] reported that 5 patients had a disastrous intraoperative recovery course, suggesting ITA hypoperfusion. All patients could be weaned from cardiopulmonary bypass only after insertion of a saphenous vein graft distal to the ITA graft or removal of the ligature from the old saphenous vein graft. Myocardial oxygen demand is markedly increased during reperfusion after anoxic arrest to compensate for oxygen deficiency. Although ITA hypoperfusion could be due to ITA spasm or technical problems at the anastomosis, the major contributing factor to inadequate ITA perfusion was considered to be a disproportion between ITA flow and myocardial demand [6, 71. Inadequate myocardial perfusion by arterial grafts is followed by ventricular dysfunction, which should decrease arterial driving pressure and increase intraventricular diastolic pressure. Both of these conditions may further decrease arterial graft flow, and a vicious cycle may start consequently. Intraaortic balloon pumping for in situ arterial grafts may not be so effective as for aortocoronary bypass grafts, because in situ arterial grafts are inferior in diastolic augmentation. Flow capacities of ITA and saphenous vein grafts, under exercise conditions, were compared using radionuclide angiocardiography. A slightly greater potential for inadequate flow in patients with ITA grafts was found, as evidenced by a small group of patients with exerciseinduced wall motion abnormalities [5]. Coronary blood flow is known to increase by 460% in response to an increase in myocardial oxygen demand [ 191. Serial changes in left ventricular function, during graded exercise, were assessed by radionuclide continuous ventricular function monitoring in patients with coronary artery bypass operation. Asymptomatic left ventricular dysfunction was found at late exercise stages in a group of patients and was considered to reflect myocardial ischemia resulting from inadequate perfusion by arterial grafts [20]. From the present study, the inferior capacity of flow through arterial grafts, compared with ascending aortaoriginated grafts, is mainly attributed to reduced diastolic pressure in the arterial grafts and is caused by anatomic characteristics. Diastolic depression might not be observed in aortocoronary grafting with a free arterial graft. The inadequacy of flow through arterial grafts might be concealed in most patients with single arterial graft and multiple vein grafts, but may be apparent in some patients with multiple arterial grafts during increased myocardial demand. It is not our purpose or intent to suggest that surgeons should not use arterial grafts. However, there may be instances in which arterial grafts originating from a systolicdominant circulation far away from the heart have some limitations in the ability to supply blood to the diastolicdominant coronary circulation. We also stress that multiple revascularization using only in situ arterial grafts can cause myocardial hypoperfusion, which aortocoronary bypass grafting may relieve due to the advantage over blood supply for coronary circulation. References 1. Loop FD, Lytle BW, Cosgrove DM, et al. Influence of the internalmammaryartery graft on 10year survival and other cardiac events. N Engl J Med 1986;314:16. 2. Singh H, Flemma RJ, Tector AJ, Lepley D Jr, Walker JA. Direct myocardial revascularization. Determinants in the choice of vein graft or internal mammary artery. Arch Surg 1973; 107~699703. 3. Flemma RJ, Singh HM, Tector AJ, Lepley D Jr, Frazier BL. Comparative hemodynamic properties of vein and mammary artery in coronary bypass operations. 1975;20:61927. 4. Hamby RI, Aintablian A, Wisoff BG, Hartstein ML. Comparative study of the postoperative flow in the saphenous vein and internal mammary artery bypass grafts. Am Heart J 1977;93:30615. 5. Kawasuji M, Tsujiguchi H, Tedoriya T, Taki J, Iwa T. Evaluation of postoperative flow capacity of internal mammary artery. J Thorac Cardiovasc Surg 1990;99:696702. 6. Jones EL, Lattouf OM, Weintraub WS. Catastrophic consequences of internal mammary artery hypoperfusion. J Thorac Cardiovasc Surg 1989;98:9027. 7. Von Segesser L, Simonet F, Meier B, Finci L, Faidutti B. Inadequate flow after internal mammarycoronary artery anastomoses. Thorac Cardiovasc Surg 1987;35:3524. 8. Carter MJ. The use of the right gastroepiploic artery in coronary artery bypass grafting. Aust N Z J Surg 1987;57 31721. 9. Lytle BW, Cosgrove DM, Ratliff NB, Loop FD. Coronary artery bypass grafting with the right gastroepiploic artery. J Thorac Cardiovasc Surg 1989;9782631. 10. Shellock FG. Transducertipped catheter. New transducertipped pressuremonitoring catheter uses fiber optics. Med Electron 1985;16:1026.
956 TEDORIYA ET AL 1993;569514 11. Barnes RJ, Comline RS, Dobson A, Drost CJ. An implantable transit time ultrasonic blood flow meter. J Physiol 1983;345: 23. 12. Von Segesser LK, Lehmann K, Turina M. Deleterious effects of shock in internal mammary artery anastomoses. Ann Thorac Surg 1989;475759. 13. Jett GK, Arcidi JM Jr, Dorsey LMA, Hatcher CR, Guyton RA. Vasoactive drug effects on blood flow in internal mammary artery and saphenous vein grafts. J Thorac Cardiovasc Surg 1987;94211. 14. Kajiya F, Tsujioka K, Ogasawara Y, et al. Analysis of flow characteristics in poststenotic regions of the human coronary artery during bypass graft surgery. Circulation 1987;76: 1092100. 15. Tedoriya T, Kawasuji M, Sakakibara N, Ueyama K, Watanabe Y. Physiological characteristics of arterial graft for coronary artery bypass surgery. Jpn J Thorac Surg 1992;45:7114. 16. Wakabayashi A, Beron E, Lou MA, Mino JY,, da Costa IA, Connolly JE. Physiological basis for the systemictocoronary artery bypass graft. Arch Surg 1970;100:179. 17. GonzalesSantos JM, Bastida E, Riesgo M, et al. Flow capacity of the human retrograde internal mammary artery: surgical considerations. 1990;50:3606. 18. Simon P, Owen A, Neumann F, et al. Immediate effects of mammary artery revascularization versus saphenous vein on global and regional myocardial function: an intraoperative echocardiographic assessment. Thorac Cardiovasc Surg 1991; 3922832. 19. Holmberg S, Serzysko W, Varnauskas E. Coronary circulation during heavy exercise in control subjects and patients with coronary heart disease. Acta Med Scand 1971;190 46580. 20. Kawasuji M, Tedoriya T, Sakakibara N, Takahashi M, Taki J, Watanabe Y. Silent left ventricular dysfunction during exercise after coronary artery bypass surgery. Eur J Cardiothorac Surg 1991;5:61&22.