Patencies of 2,127 Arterial to Coronary Conduits Over 15 Years

Transcription

1 Patencies of 2,127 Arterial to Coronary Conduits Over 15 Years James Tatoulis, FRACS, Brian F. Buxton, FRACS, and John A. Fuller, FRACP Royal Melbourne Hospital and Epworth Hospital, University of Melbourne, Melbourne, Victoria, Australia Background. Use of arterial grafts in coronary surgery is based on the excellent patency of the left internal thoracic artery (LITA) and an expectation that other arterial grafts right internal thoracic artery (RITA) and radial artery (RA) will give similar patencies, superior to saphenous vein. We examined patencies of arterial grafts in a practice with extensive use for more than 15 years. Methods. Consecutive postoperative angiograms of 2,127 arterial/coronary conduits were evaluated. Angiograms were performed for cardiac symptoms. Assessment was by two observers. String signs were considered as occlusions. Results. There were 2,127 arterial conduits. Overall patencies were as follows: LITA, 96.4% (1,296 of 1,345); RITA, 88.3% (534 of 605); aortocoronary RA, 89.3% (158 of 177). The LITA patency to the left anterior descending artery was 97.1% (1,131 of 1,165); to the obtuse marginal artery it was 91.7% (165 of 180; p 0.01). The RITA pedicled graft patency was 86% (275 of 321) compared with free RITA, 91% (259 of 284; p not significant). For RITA there was a hierarchy of patency for coronary territory grafted (left anterior descending artery best, right coronary/posterior descending artery worst) and for degree of coronary stenosis: if stenosis was less than 60%, patency was 65% (47 of 72); if stenosis was more than 60%, patency was 90.9% (485 of 533; p ). Similarly for the radial artery there was higher patency with greater coronary stenosis. The LITA patency at 5 years was 98%, at 10 years it was 95%, and at 15 years it was 88%. The RITA patency at 5 years was 96%, at 10 years it was 81%, and at 15 years it was 65%. The radial artery patency at 1 year was 96% and at 4 years it was 89%. For 3,714 vein grafts also studied overall patency was 61% (2,266 of 3,214) with patencies of 95% at 5 years, 71% at 10 years, and 32% at 15 years. Conclusions. Excellent long-term patencies of arterial grafts are noted, superior to those of vein grafts. Patencies were influenced by conduit, by distribution, and by coronary artery stenosis. (Ann Thorac Surg 2004;77:93 101) 2004 by The Society of Thoracic Surgeons The development of occlusive atherothrombotic disease in coronary saphenous vein grafts (SVG) has reduced long-term SVG patency and is inferior to that of arterial conduits in coronary artery bypass graft (CABG) surgery [1 3]. The preferential use of arterial conduits has been associated with improved clinical results, which may be due to better patencies than for SVG [4 8]. We have progressively increased the use of arterial grafts initially in 1985 with the left internal thoracic artery (LITA) followed in 1986 by bilateral internal thoracic arteries and then in 1996 by the routine use of the radial artery (RA) in addition to the internal thoracic arteries [9]. Our objective was to achieve total arterial coronary revascularization if possible as the clinical results were as good or better than where extensive saphenous vein grafting was used with the additional expectation of superior long-term clinical and patency results, which would include reduced need for coronary reoperation. In our practice total arterial coronary revascularization is possible in more than 90% of all CABG with excellent results [9]. Presented at the Thirty-ninth Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31 Feb 2, Address reprint requests to Dr Tatoulis, Suite 28, Private Medical Centre, Royal Melbourne Hospital, Parkville, Victoria, 3050 Australia; james.tatoulis@mh.org.au. Although there is much documentation of clinical results in CABG there is relatively little information on long-term arterial/coronary conduit patencies. The purpose of this study was to examine the long-term patencies of a large number of arterial grafts in an experience that spans 15 years and to document the anatomic and physiologic behavior of the common arterial conduits. Additionally we wished to evaluate any specific anatomic (conduit or target vessel) factors that influenced longterm arterial graft patency and to document the modes of failure of arterial grafts in the coronary circulation. Material and Methods We studied 2,127 consecutive arterial/coronary conduit angiograms performed from January 1987 to December 2001 and this cohort comprises 1,408 patients undergoing postoperative coronary graft angiography. For each patient 1.5 arterial coronary conduits were studied. All arterial/coronary conduit angiograms in this cohort were performed for postoperative cardiac symptoms. In this study only patients who had coronary graft angiograms for possible cardiac symptoms were considered. These symptoms included recurrent typical angina, atypical chest pain, new or persistent shortness of breath, or presentation with a new myocardial infarction. Asymp by The Society of Thoracic Surgeons /04/$30.00 Published by Elsevier Inc doi: /s (03)

2 94 TATOULIS ET AL Ann Thorac Surg ARTERIAL TO CORONARY CONDUITS 2004;77: tomatic patients did not have coronary angiograms. There was no routine postoperative coronary angiography policy with respect to these patients. During this period there were 10,407 CABG operations in which one or more arterial grafts were used. The rate of total arterial coronary revascularization has risen from 30% in 1995 to 90% since Arterial Graft Procurement and Antispasm Prophylaxis The LITA was initially harvested as a narrow pedicle with low-power cautery. Intercostal branches were divided by scissors between titanium clips. Since 1990 the LITA has been harvested with only the medial vein in a skeletonized or semiskeletonized manner. Inferiorly the dissection always extended to the internal mammary artery bifurcation and proximally as far as possible, dividing all branches. The RITA was harvested in an identical fashion. For both internal mammary arteries, 2 ml of intraluminal papaverine solution (1 mg/ml in heparinized ringers lactate, ph 7.2 at 37 C) has been used. The end is clipped and the internal mammary artery pedicle enveloped in a papaverine solution soaked gauze and left pulsating in situ until use. Radial artery procurement followed the same principles. The radial artery was harvested together with its venae commitantes and stored in a papaverine solution papaverine 1 mg/ml in heparinized arterial blood at room temperature until use. Radial arteries with significant calcification were not harvested. Operative details have been previously published [9 11]. All patients received intravenous nitroglycerine (30 g to 100 g/min intraoperatively and for the first 24 hours) and all received enteric-coated aspirin 100 mg daily commencing on the first postoperative day and continuing indefinitely. The calcium-channel blocker amlodipine (5 mg orally daily for 6 months) was used routinely when an radial artery was used [9 11]. Graft Strategy: Deployment of Arterial/Coronary Conduits The LITA was used as a pedicled graft almost exclusively to the left anterior descending artery (LAD) system 13% of pedicled LITA grafts were placed to the circumflex marginal artery (CxOM) system. The RITA was used as a pedicled graft if possible. However it was readily used as or converted to a free graft to reach a more favorable site for anastomosis and to avoid stretching. The RITA was more commonly grafted to the right coronary artery/ posterior descending artery, followed by the CxOM, and least to the LAD. The RITA when used as a free graft was almost always from the aorta (seldom from the LITA and never from an SVG). A 3.5-mm opening was created in the ascending aorta with a pumch and the anastomosis constructed with continuous 6-0 polypropylene. The proximal end of the free RITA is usually 3 mm internal diameter and has a robust wall. A vein patch on the aorta to facilitate proximal RITA anastomosis has only been used on one occasion. The radial artery grafts were usually deployed to the circumflex or posterior descending artery and left ventricular branches of the RCA. The LITA-Y-radial artery composite grafts were infrequently performed and not included in this study. The radial artery angiograms in this study were aortocoronary only. Sequential grafting with arterial conduits was rarely used through this experience, extremely rarely before 1997 and seldom from 1997 to 2001 (occasionally with radial arteries). Postoperative angiograms with sequential grafts were excluded from the study. Supplementary SVG were used as required to achieve complete myocardial revascularization (mean of 3.1 distal anastomoses per patient). Since 1999, fewer than 10% of patients have received an SVG. The right gastroepiploic artery, inferior epigastric artery, and ulnar artery were infrequently used. In this cohort all CABG were performed using cardiopulmonary bypass, aortic clamping, and cardioplegic arrest. Since 1990 a single cross clamp and combined antegrade and retrograde blood cardioplegia has been used [9 11]. Postoperative Angiography Angiograms were performed under local anesthesia usually by the common femoral artery approach within a cardiac catheterization laboratory. All patients gave informed consent to reangiography and study of their coronary conduits. All angiograms were reported by two observers independently. Conduit/graft failure was defined as total occlusion, a string sign (conduit diameter 1 mm), or a stenosis of 80% or greater anywhere within the conduit or at the anastomosis. An analysis was undertaken to relate arterial/coronary conduit patency to time, conduit type, free or pedicled, coronary territory of conduit deployment, and native coronary artery stenosis. The LITA-radial artery Y- pedicled grafts were excluded from this analysis as they required detailed consideration of the multiple LITA segments, the proximal internal mammary artery/radial artery anastomosis, each subsequent segment of radial artery, and anastomosis. It was thought that this would confound the patency data relating to simple aortocoronary radial artery conduits. Data Collection and Analysis All data (operative and angiographic) were collected prospectively and entered into a computer database program. Intraoperative data were entered at the time of operations. Remote angiographic data were entered at the time of the angiogram report. Values are reported as the mean SD. Percentages are given where appropriate. Statistical analysis was with the Statistical Package for Social Sciences (SPSS-PC plus). The 2 test was used for categorical variables. Conduit survival curves were calculated using the Kaplan-Meier method and compared with the log-rank test. A p value of less than 0.05 was considered significant. Results In all 2,127 arterial conduits were studied. The mean interval from operation to postoperative angiographic study was months (range, 1 to 212). Time to

3 Ann Thorac Surg TATOULIS ET AL 2004;77: ARTERIAL TO CORONARY CONDUITS Table 1. Overall Arterial Conduit Patencies Conduit Number Time (months) Failed Conduits Patency Number Percent LITA 1, / % RITA / Ao RA / All conduits 2, / Ao RA aortocoronary radial artery; LITA left internal thoracic artery; RITA right internal thoracic artery; Time mean time (months) to postoperative angiogram. postoperative angiography was longer for LITA (79.3 months) and for RITA (81.9 months) compared with that for radial artery (26.6 months). The mean interval from operation to angiography for the 1,345 LITA grafts was 79.3 ( 46.4) months and for the 605 RITA grafts it was 81.9 ( 43.7) months. The majority of the LITA angiograms (887 of 1,345; 66%) were performed 5 or more years postoperatively; for the RITA angiograms also the majority (421 of 605; 69%) were performed 5 or more years postoperatively. For the LITAs for most 1-year intervals approximately 100 graft studies were performed and this was steady from year 1 postoperative to year 11 postoperative. For the RITA angiograms an average of 50 were performed on a very even basis from year 2 postoperative to year 11 postoperative. The radial artery postoperative angiograms are in a much shorter time frame, with an average of 30 being performed per annum again on a fairly regular basis. The specific details of yearly angiogram numbers for each of the three types of arterial conduits are given in Appendixes 1, 2, and 3. For all arterial grafts studied (n 2,127) irrespective of type there were 139 graft failures. Overall patency for all arterial conduits was 93.5% (1,988 of 2,127) 76 months postoperatively. Table 2. Patency of Arterial/Coronary Conduits by Territory LAD CxOM RCA/ PDA p Value LITA 97.1% 91.7% Number 1131/ /180 RITA (attached) 94.6% 84.9% 79.5% a Number 106/112 45/53 124/156 RITA (free) 96.2% 91.9% 88.8% Number 25/26 147/160 87/98 RITA (all) 94.9% 90.1% 83.1% b Number 131/ / /254 RA 87.1% 91.6% 88.2% Number 27/31 87/95 45/51 SVG 60.2% 61.2% 61.6% Number 728/ / /1208 a p Value between each group. b p Value RITA-LAD and RITA-RCA/ PDA groups. CxOM circumflex artery; LAD left anterior descending artery; LITA left internal thoracic artery; RA radial artery; RCA/PDA right coronary/posterior descending artery; RITA right internal thoracic artery; SVG saphenous vein graft. LITA Conduit Patencies The coronary territory influenced LITA patency. When the LITA was anastomosed to the LAD territory, patency was 97% (1,131 of 1,165). For LITA to the circumflex marginal, patency was less, at 92% (165 of 180; p 0.012). The probability of LITA patency to be left anterior descending is 97% (1,131 of 1,165) at a mean of 80 months and overall LITA patency to the circumflex marginal is 92% (165 of 180) at a mean of 80 months. More specifically LITA to LAD patency is 99%, 97%, and 93% at 5, 10, and 15 years respectively whereas LITA to circumflex marginal patencies are 97%, 92%, and 89% at 5, 8, and 10 years respectively. RITA Conduit Patencies Overall patency for pedicled (in situ) RITA grafts was 86% (275 of 321). By comparison, patency for free RITA grafts was 91% (259 of 284; p 0.564). Coronary artery territory influenced RITA patency. There was a hierarchy of patency, with best patency to the LAD then the CxOM, and lastly the right coronary artery/posterior descending artery territory. For RITA to LAD patency was 95% (131 of 138). For RITA to CxOM patency was 90% (192 of 213). For RITA to the right coronary artery/posterior descending artery patency was 83% (211 of 254; p 0.01 between each group). 95 Overall Patencies by Conduit The LITA had the best patency, then the RITA, and finally the aortocoronary radial artery. Of the 1,345 LITA conduits studied there were 49 LITA failures. Overall patency was 96.4% (1,296 of 1,345). For the 605 RITA conduits there were 71 RITA failures. Overall RITA patency was 88.3% (534 of 605). For the 177 aortocoronary radial artery conduits there were 19 failures. Overall aortocoronary radial artery patency was 89.3% (158 of 177). These results are summarized in Table 1. Aortocoronary Radial Artery Conduit Patencies Patencies to the different territories were similar. Coronary territory did not appear to influence patency of the aortocoronary radial artery. Radial artery patency to the LAD was 87% (27 of 31), to the circumflex marginal it was 92% (87 of 95), and to the right coronary artery/posterior descending artery territory it was 88% (45 of 51; p 0.392). Patencies of arterial conduits by territory are detailed in Table 2. In comparison with the LITA and RITA grafts to the LAD, radial artery grafts were clearly inferior. Radial artery to LAD grafts were used when the LITA was deployed to the CxOM if it was a more important vessel, when the LAD was diffusely diseased and had to be grafted distally, and when the LITA could not reach the intended site for anastomosis (especially in emphysematous patients with inferior displacement of the heart), when there was LITA trauma or dissection or chest wall radiotherapy, and in descending/abdominal aortic occlusion, where the internal mammary arteries formed the main collaterals to the lower body. For the CxOM all three arterial conduits had similar patencies (LITA 92%,

4 96 TATOULIS ET AL Ann Thorac Surg ARTERIAL TO CORONARY CONDUITS 2004;77: RITA 91%, radial artery 92%) although time to study for radial artery grafts was significantly less. For the right coronary artery/posterior descending artery territory, radial artery patency (88%) was better than that for RITA (83%). However there are difficulties in this comparison because of different mean times to angiography (27 months for radial artery versus 82 months for RITA). Pedicled Internal Thoracic Artery Grafts to the LAD Patencies for both pedicled (in situ) right internal thoracic artery and left internal thoracic artery grafts to the LAD were excellent. For the pedicled LITA to LAD, patency was 97% (11,31 of 1,345) and for the pedicled RITA to LAD, patency was 95% (131 of 138; p 0.821). Cumulating all the conduits, the probability of having a patent conduit according to the territories are as follows: for the left anterior descending 97% (1,289 of 1,334); for the circumflex 91% (444 of 488); and for the right coronary artery/posterior descending artery 84% (256 of 305) at a mean of 76 months postoperative. Table 3. Target Coronary Artery Stenosis Influence on Patency Conduit Target Vessel Stenosis 60% 60% p Value LITA patency 92% 97% Number 113/ /1222 RITA patency 65% 91% Number 44/66 490/539 80% 80% RA patency 83% 92% Number 50/60 108/117 60% 60% SVG patency 56% 62% Number 194/ /3367 LITA left internal thoracic artery; RA radial artery; RITA right internal thoracic artery; SVG saphenous vein. Native Coronary Artery Stenosis: Influence on Patency There was a significant influence by the degree of native coronary artery stenosis on arterial conduit patency. The more severe the native coronary artery stenosis, the higher the patency. The threshold stenosis for this effect was a stenosis of 60% for LITA and RITA. Radial artery conduits were more sensitive; the threshold stenosis was 80%. Although pedicled LITA graft patencies followed the same patterns, the effects on LITA patency were less severe. The results are summarized in Table 3. The poorest patency was for the RITA to coronary arteries with a stenosis of less than 60% (usually pedicled RITA to large right coronary arteries; Fig 1). Patency Over Time For all arterial conduits, there was a steady decline of patency over time. This was less marked for the LITA, followed by the RITA and then the radial artery (Fig 2). The LITA patencies at 5, 10, and 15 years were 98%, 95%, and 88% respectively. The RITA patencies at 5, 10, and 15 years were 96%, 81%, and 65% (p for 10- and 15-year comparisons). Radial artery patency at 1 year was 96% and at 4 years it was 89%. Comparison of Saphenous Vein Conduits During the study period 3,714 SVG were also studied (mean time to angiography months). There were 1,448 SVG failures. Overall patency was 61% (2,266 of 3,714). For SVG patency was not affected by territory: LAD, 60.2% (728 of 1,210), CxOM, 61.2% (793 of 1,296), and right coronary artery/posterior descending artery, 61.6% (745 of 1,208; p 0.688). Additionally SVG patency was less affected by the degree of native coronary artery stenosis. When the stenosis was more than 60%, SVG patency was 61.5% (2,072 of 3,367); when the stenosis was less than 60%, patency was 55.9% (194 of 347; p 0.030). The SVG patencies at 5, 10, and 15 years were 95%, 61%, and 32% respectively (p 0.001, SVG versus RITA, 10- and 15-year data; Fig 2). Arterial Conduit Failures: Qualitative Observations When arterial grafts failed, there were three modes: total occlusion; a string sign where the length of the graft could be seen from its origin to the target coronary artery, however the graft diameter was less than 1 mm and deemed nonfunctional; or a discrete stenosis (Figs 3 and 4).There were no intraluminal irregularities suggesting atheroma or thrombus in contrast to the changes seen in patent diseased SVG. Patent arterial grafts were uniform and had larger diameters than the target coronary arteries (Fig 5). The incidence of arterial/coronary conduit spasm was exceedingly low. When spasm occurred it was usually in radial artery (4 cases) weeks to months postoperatively, localized more than 1 cm, and Fig 1. Arterial/coronary conduit patency: effect of target vessel stenosis. For the left internal thoracic artery (LITA) and right internal thoracic artery (RITA), patencies when grafted to coronary vessels of less than 60% stenosis and more than 60% stenosis are compared (p 0.01, p respectively). For the radial artery (RA) differences in patencies are noted for target vessel stenosis of less than 80% and more than 80% (p 0.01).

5 Ann Thorac Surg TATOULIS ET AL 2004;77: ARTERIAL TO CORONARY CONDUITS 97 Fig 2. Patency of coronary conduits over time for left internal thoracic artery (diamonds, n 1,345), right internal thoracic artery (RITA [boxes, n 605]), radial artery (triangles, n 177), and saphenous vein grafts (VG [circles, n 3,714]). There are significant patency differences between each group at 10 years (p 0.001) and also at 15 years (p 0.001). Patency of RITA at 15 years is twice that for VG. resolved angiographically with intraluminal nitroglycerin. Comment Saphenous vein bypass grafts in the coronary circulation, although providing excellent short and medium-term Fig 4. (A, B) Two views of a discrete, severe stenosis (S) in an otherwise normal radial artery (RA) conduit 4 years postoperatively. The cause of the stenosis was unknown and it was treated successfully by angioplasty and stent. Fig 3. A string sign in the left internal thoracic artery (LITA) because of competitive flow. The LITA had been previously anastomosed to a large (2.5 mm) left anterior descending artery that had a diameter stenosis of 40%. results, have been disappointing in the longer term resulting in recurrence of cardiac events, hospital readmissions, and need for reintervention. Beyond 5 years SVGs are vulnerable to intimal hyperplasia and atherothrombotic occlusive disease [1 3]. By comparison arterial grafts particularly the LITA and also the RITA have been associated with better patencies and improved clinical results including fewer cardiac events and a greater life expectancy [4 8]. The results with respect to the radial artery are promising [10 15] and comparable with or better than when SVGs are used, particularly as a second conduit of choice after the internal mammary artery. Since 1985 we have incrementally used more arterial grafts to achieve better long-term clinical outcomes. That has resulted in a large experience with arterial grafts and with total arterial coronary revascularization [9]. This study of 2,127 arterial/coronary conduits at a mean interval of almost 80 months represents one of the largest late angiographic studies of arterial/coronary conduits. The large numbers allow greater confidence in interpretation of patterns of conduit patency or failure. The angiograms were performed in a symptomatic group of patients and hence the observed results may be biased toward poorer patencies. Despite this the patencies for arterial grafts are excellent and at 10 and 15 years are clearly superior to SVG.

6 98 TATOULIS ET AL Ann Thorac Surg ARTERIAL TO CORONARY CONDUITS 2004;77: Fig 5. The left internal thoracic artery (LITA) displays typical characteristics of a patent arterial to coronary conduit 6 years postoperatively. It is of uniform diameter and larger than the coronary vessel it supplies, the obtuse marginal (OM) artery. Note the absence of any luminal irregularity. For the LITA this study confirms the excellent patencies noted by Loop and associates [4] and others [16 18]. Nevertheless the pedicled LITA displayed a slightly lower patency when grafted to the CxOM. This may be due to greater technical difficulty, stretching of the LITA pedicle to reach an inferior marginal, and smaller less optimal target vessels with a more limited runoff than the LAD. Coronary territory also influenced RITA patency. A hierarchy of patency was observed best to the LAD then the CxOM and finally the right coronary artery/ posterior descending artery. Hence there may be benefits of RITA grafting to the left sided coronary arteries [19]. Patency of the RITA to the LAD was identical to that of the LITA to the LAD indicating that equivalent biological conduits to the same vascular bed produce identical results. The superior patency of the RITA when placed to the LAD by comparison to the other territories may be explained by shorter graft length, technical ease, and better runoff. In this large series of RITA angiograms there were similar numbers of pedicled and free RITAs. Overall patencies of these groups were similar. Within each of the pedicled and the free RITA groups, again a hierarchy of patency was observed, the poorest patency being to the right coronary artery/posterior descending artery, indicating that the coronary territory and its vascular bed is a significant determinant of long-term patency. With respect to the right coronary artery/posterior descending artery system a pedicled RITA may provide limitation as to how distal it will reach sometimes resulting in construction of the anastomosis in a suboptimal area of wall thickening. Additionally the RCA is notorious for the development and progression of further disease distal to an anastomosis (usually near the crux). Under these circumstances if the RITA is to be used the patient may be best served by a free RITA graft to the PDA beyond proximal wall disease. The routine procurement of the internal mammary arteries as skeletonized grafts ensures a further 2 to 3 cm in length, enabling more distal anastomosis if desirable. Although initially there was concern that internal mammary artery skeletonization may lead to trauma or disruption of the vascular supply of the internal mammary artery wall, the clinical and early angiographic results in skeletonized internal mammary artery are excellent [20]. Use of the radial artery as the second conduit of choice after the internal mammary artery is relatively recent and increasingly popular. In this study radial artery patency although inferior to that of internal mammary arteries was excellent. Again there was a tendency for lower patency when the radial artery was anastomosed to the right coronary artery/posterior descending artery. Patency information for the radial artery is still being gathered but this experience concurs with other reports [11, 12, 14, 15, 21, 22]. To date the radial artery patencies are excellent at 96% at 1 year and 90% at 27 months. Another 5 years of data are required before meaningful comparisons to the internal mammary artery can be made. Native coronary artery stenosis had a major influence on arterial conduit patency. This study reconfirms earlier observations that for arterial grafts the best patency is achieved when placed to tightly stenosed or occluded coronary vessels [12, 14, 21, 22]. All three conduits (LITA, RITA, and radial artery) whether pedicled or free behaved similarly. The pedicled LITA was the most versatile and least sensitive. By contrast the radial artery was the most sensitive, patency being significantly affected if native coronary artery stenosis was less than 80% (by comparison with the internal mammary arteries, native coronary artery stenosis 60%). The greater sensitivity of the radial artery may be due to its thicker muscular wall [23]. Native coronary artery stenosis influences right gastroepiploic artery patency in a similar manner [24, 25]. The lowest patencies were achieved when the RITA (attached or free) or the radial artery were anastomosed to the right coronary artery with less than 60% stenosis implicating a number of etiologic factors such as wall disease, progression of distal lesions, competitive flow, and vascular remodeling [23]. Also the concept of a 60% stenosis is relative. A 60% stenosis in a 5-mm right coronary artery will result in different (greater) native coronary flow and competitive flow by comparison to a 60% stenosis in a 2-mm LAD or CxOM. The large dominant right coronary artery with a moderate (50% to 60%) stenosis presents a dilemma. Some authors argue that an excellent SVG may be a more appropriate conduit in that setting [25]. Rapamycin drugeluting stents may also have a role. Patencies noted in this study are similar to, confirm, and add to the results published for LITA [4, 16 18, 20],

7 Ann Thorac Surg TATOULIS ET AL 2004;77: ARTERIAL TO CORONARY CONDUITS not seen angiographically (nor clinically at redo operations) in arterial conduits. Implications are that arterial conduits are not susceptible to abrupt closure (in contrast to a severely diseased and stenosed SVG) and in a reoperation, an arterial graft would not be a hazardous source of accidental atheroembolism making coronary reoperations (if such had to be undertaken) safer. Ten years postoperatively we should expect arterial/conduit patencies of between 80% (radial artery, RITA) and 95% (LITA) and at 15 years, patencies of 65% to 90%, which are clearly superior to those of SVG. The concept of improved long-term clinical outcomes and prognosis when increased numbers of arterial grafts are used is supported by many authors [4 9]. Based on the present data optimal patencies (and clinical results) are achieved by deployment of the LITA to the LAD, the RITA (pedicled if it will comfortably reach or free) to the CxOM and the radial artery to the posterior descending artery. 99 Fig 6. Distal disease (DIS) developed in an obtuse marginal (OM) artery 4.5 years postoperatively. The OM had been large, normal, and uniform at the time of surgery. (RA radial artery.) for RITA [15, 16, 19], and for radial artery [11, 12 15, 21, 22]. With arterial coronary conduits it is hoped that once technical and competitive flow factors are overcome that constant patency would be maintained over the long term. However there is still some late graft failure. Possible explanations include a conduit that may have occluded at year 2 may not be discovered until year 10, when for some additional reason the patient becomes symptomatic and an angiogram is performed. The progress of coronary lesions distal to the anastomosis, resulting in poor runoff and graft failure (Fig 6). We have observed patent grafts to subsequently occlude on serial angiograms. Additionally intrinsic biologic factors affecting myocytes, fibroblasts, and endothelial cells in the arterial conduit wall may result in long-term changes [23]. Some patients can display an intense inflammatory reaction around the conduit (seen at reoperation), which may result in longer term perigraft fibrosis and conduit failure. The anatomic modes of arterial/conduit failure are distinct and different from the intimal hyperplasia and occlusive atherothrombotic disease seen in SVG. Arterial conduits display total occlusions, string signs, or localized stenoses. The total occlusions may be due to early technical factors. The string signs are definitely seen in the setting of competitive flow and may represent an adaptation to low flow demand. The long-term fate of string sign arterial grafts remains unknown and controversial [21 23] whereas localized stenoses are possibly associated with either specific sites of trauma or areas of localized wall disease in the arterial conduit occasional plaques of atheroma in the internal mammary artery, or wall calcification in the radial artery which then progress to a stenosis. Unlike SVG occlusive atherothrombotic changes are Limitations of the Study Limitations include that it is an observational study of symptomatic patients and possibly biased toward lower patencies. The exact time of the conduit occlusions is not known. It is assumed that they occurred at the time of angiography. However all occlusions must have occurred a variable period of time before angiography. The angiographic cohort represents less than 15% of all patients undergoing CABG with arterial grafts during this time frame. Reporting of minor abnormalities in grafts or at anastomotic points is problematic. Unless the changes were severe (stenosis 80%), conduits were reported as normal. The radial artery follow-up is short and LITA/ radial artery Y grafts have been excluded. Prior reports of LITA/radial artery Y grafts have documented excellent LITA to LAD and radial artery to OM patencies but poor patencies of the distal limb of the radial artery Y graft to the posterior descending artery. Patency in the Y radial artery limb is also significantly influenced by native coronary stenosis and competitive flow [21, 22]. In conclusion in this study of a large number of late postoperative arterial/coronary conduit angiograms excellent long-term patencies of arterial conduits are documented and are superior to those of saphenous vein grafts. Patencies were influenced by conduit type, coronary artery territory to which they were applied, and by the degree of target native coronary artery stenosis. References 1. Grondin C, Campeau L, Lesperance J, Enjalbert M, Bourassa M. Comparison of late changes in internal mammary artery and saphenous vein grafts in two consecutive series of patients 10 years after operation. Circulation 1984;70(Suppl 1): Lytle BW, Loop FD, Cosgrove DM, Ratliff NB, Easley K, Taylor PC. Long-term (5 to 12 years) serial studies of internal mammary artery and saphenous vein coronary bypass grafts. J Thorac Cardiovasc Surg 1985;89: FitzGibbon GM, Cafka HP, Leach AJ, Keon WJ, Hooper D, Burton JR. Coronary bypass graft fate and patient outcome: angiographic follow-up of 5065 grafts related to survival and

8 100 TATOULIS ET AL Ann Thorac Surg ARTERIAL TO CORONARY CONDUITS 2004;77: re-operation in 1388 patients over 25 years. J Am Coll Cardiol 1996;28: Loop FD, Lytle BW, Cosgrove DM, et al. Influence of the internal-mammary-artery graft on 10-year survival and other cardiac events. N Engl J Med 1986;314: Lytle BW, Blackstone EH, Loop FD, et al. Two internal thoracic artery grafts are better than one. J Thorac Cardiovasc Surg 1999;117: Schmidt SE, Jones JW, Thornby JL, Miller CC III, Beall AC JR. Improved survival with multiple left sided internal thoracic artery grafts. Ann Thorac Surg 1997;64: Buxton BF, Komeda M, Fuller JA, Gordon I. Bilateral internal thoracic artery grafting may improve outcome of coronary artery surgery. Risk-adjusted survival. Circulation 1998; 98(Suppl II):II Pick A, Orszulak T, Anderson B, Schaff H. Single versus bilateral internal mammary artery grafts: 10-year outcome analysis. Ann Thorac Surg 1997;64: Tatoulis J, Buxton BF, Fuller JA, Royse AG. Total arterial coronary revascularization: techniques and results in 3,220 patients. Ann Thorac Surg 1999;68: Tatoulis J, Buxton BF, Fuller JA, Royse AG. The radial artery as a graft for coronary revascularization: techniques and follow-up. In: Karp R, Lacks H, Weschler A, eds. Advances in cardiac surgery. Vol 11. St Louis: Mosby, 1999: Tatoulis J, Buxton BF, Fuller JA, et al. The radial artery in coronary surgery: a 5 year experience-clinical and angiographic results. Ann Thorac Surg 2002;73: Acar C, Ramsheyi A, Pagny JT, et al. The radial artery for coronary artery bypass grafting: clinical and angiographic results at five years. J Thorac Cardiovasc Surg 1998;116: Moran SV, Baeza R, Guarda E, et al. Predictors of radial artery patency for coronary artery bypass grafting. Ann Thorac Surg 2001;72: Iacò AL, Teodori G, Di Giammarco G, et al. Radial artery for myocardial revascularization: long-term clinical and angiographic results. Ann Thorac Surg 2001;72: Parolari A, Rubini P, Alamanni F, et al. The radial artery: which place in coronary operation? Ann Thorac Surg 2000; 69: Dion R, Glineur D, Derouck R, et al. Long-term clinical and angiographic follow-up of sequential internal thoracic artery grafting. Eur J Cardiothorac Surg 2000;17: Ascione R, Underwood MJ, Lloyd CT, Jeremy JY, Bryan AJ, Angelini GD. Clinical and angiographic outcome of different surgical strategies of bilateral internal mammary artery grafting. Ann Thorac Surg 2001;72: Mack MJ, Osborne JA, Shennib H. Arterial graft patency in coronary artery bypass grafting: what do we really know? Ann Thorac Surg 1998;66: Buxton BF, Ruengsakulrach P, Fuller J, Rosalion A, Reid CM, Tatoulis J. The right internal thoracic artery graft-benefits of grafting the left coronary system and native vessels with a high grade stenosis. Eur J Cardiothorac Surg 2000;18: Calafiore AM, Vitolla G, Iacò AL, et al. Bilateral internal mammary artery grafting: midterm results of pedicled versus skeletonized conduits. Ann Thorac Surg 1999;67: Royse AG, Royse CF, Tatoulis J, et al. Post-operative radial artery angiography for coronary artery bypass surgery. Eur J Cardiothorac Surg 2000;17: Maniar HS, Sundt TM, Barner HB, et al. Effect of target stenosis and location on radial artery graft patency. J Thorac Cardiovasc Surg 2002;123: Barner HB. Remodeling of arterial conduits in coronary grafting. Ann Thorac Surg 2002;73: Ochi M, Hatori N, Fuji M, et al. Limited flow capacity of the right gastroepiploic artery graft: postoperative echocardiographic and angiographic evaluation. Ann Thorac Surg 2001;71: Suma H, Isomura T, Horii T, Sato T. Late angiographic result of using the right gastroepiploic artery as a graft. J Thorac Cardiovasc Surg 2000;120: Appendix 1 Angiograms by Yearly Intervals: Left Internal Thoracic Artery Time to Postoperative Angiogram Greater than...(years) Number of Angiograms Angiograms in the Specific Yearly Interval Appendix 2 Angiograms by Yearly Intervals: Right Internal Thoracic Artery Time to Postoperative Angiogram Greater than...(years) Number of Angiograms Angiograms in the Specific Yearly Interval Appendix 3 Angiograms by Yearly Intervals: Radial Artery Time to Postoperative Angiogram Greater than...(years) Number of Angiograms Angiograms in the Specific Yearly Interval

9 Ann Thorac Surg TATOULIS ET AL 2004;77: ARTERIAL TO CORONARY CONDUITS DISCUSSION DR SOICHIRO KITAMURA (Osaka, Japan): Doctor Baumgartner, Dr Murray, members, and guests. This is a slide showing the chronological patency rates of various bypass grafts utilized by the authors and depicted from the abstract. First, the initial 1- to 5-year patency rates were more than 95% regardless of the nature of the graft, which means their excellent indication and technique for a modern bypass operation. The difference in patency rate with different grafts becomes apparent at the late period, usually after 5 years. The left ITA when anastomosed to the LAD had the best patency rate, which is a common finding among the previous studies. The high late attrition rate of the right ITA in the series of Dr Tatouli s is somewhat peculiar, most probably due to quality and nature of the recipient target vessels, for example the right coronary artery. But to our mind there is essentially no difference in quality between the right and left ITA. The 5-year patency of 95% for the vein graft is amazing but the high late closure rate is universal owing to progressive atherosclerosis of the graft. The relatively high closure rate of 21% for the radial artery graft at 5 years, although excellent at 1 year, is somewhat disappointing. This rate is probably due to the high incidence of string phenomenon of the radial artery graft when anastomosed to the artery with physiologically insignificant stenosis. The radial artery grafts tends to undergo thinning more frequently than the ITA and the vein graft. The authors considered thinned graft or string graft as occluded but that is a very controversial point. Arterial grafts with functioning endothelium undergo arterial remodeling to maintain shear stress to the endothelial wall, which enhances the production of nitric oxide when flow through it decreases. Suppose the radial artery and ITA endothelial cells function equally to the same level of shear stress against arterial remodeling, flow of the radial artery graft must be eight times larger than that of the ITA graft after bypass operation. In other words, after CABG when graft flow is the same through the radial artery graft and ITA graft, remodeling or string is several times more common for the radial artery graft than for the ITA graft. So I would say that radial artery graft may be the suitable graft material for sequential grafting to increase the flow through it. Vein graft does not respond well to the shear stress change. The diameter of arterial conduits reduces to the various degree from thinning or a string sign, to oscillating flow phenomenon, no-flow patency, and eventual thrombosis. The clear-cut separation of various levels of arterial remodel is difficult. This slide shows our data published in 1992 (J Thorac Cardiovasc Surg 1992;104:1532 8). When the degree of coronary stenosis is less than 60% to 70%, the ITA graft size begins to reduce from the original diameter 1.5 mm but this remodeling process can reverse when flow requirement increases if not thrombosed. This slide shows the angiographic demonstration of no-flow anatomical patency on the ITA graft. In the left panel you see the ITA graft looks closed according to Dr Tatoulis criteria but when the recipient LAD was temporarily occluded with a PCI balloon, then the ITA showed apparent flow to the LAD as shown in the middle panel. No flow or thinned grafts have a potential to redilate and grow in size later depending upon the flow requirement as shown in the right panel. Angiographic judgment of patency is somewhat equivocal with regard to arterial grafts, and thus angiographic indications for and the planning and execution of a reoperation are also equivocal. The endotheliumdependent arterial remodeling process is propitious in the circumstance of increased conduit flow but unfavorable in the 101 circumstance of reduced flow. We may need specific medications to counteract the remodeling process of the radial artery graft and other arterial grafts as well. My question for Dr Tatoulis is threefold. First, you mentioned that the thinned artery graft was judged as closed. Then what was the rate of string phenomenon contributing to your closure rate of each arterial graft? If you consider that the string effect of the arterial graft is undesirable and should be counted as graft closure, who or what would be responsible for this undesirable result, cardiologists or surgeons who made an indication or recipient artery or graft itself? How often do you replace the thinned artery graft as a cause of unfavorable results? Second, can surgeons overcome this shortcoming of a live arterial conduit? When we place the graft to the coronary artery with 50% stenosis, probably we better put this anastomosis side to side in sequential fashion, ending with a more stenotic artery, or use a radial artery as a stem graft before bifurcated to multiple target anastomoses. Do you agree with this strategy? How do you speculate on the late fate of the radial artery graft over 10 years? Last, about antispasm or anticoagulation therapy for arterial conduits: what kind of and how long do you give the patient those medications, particularly for the patient with a radial artery graft? DR TATOULIS: Thank you, Dr Kitamura, for that discussion and the questions. Briefly to answer the complex issues, regarding string signs, we are still exploring what this means. I think it probably means a spectrum of things. In the pedicled grafts like the left internal thoracic artery graft, a string sign may not represent failure; it may represent remodeling. And with occlusion of the native LAD, perhaps the left internal thoracic artery can open up again as has been demonstrated in the literature. However for free grafts like radial artery grafts and free right internal thoracic artery grafts, I believe that if the string sign has been present for more than a few weeks or months it would be permanent. We have reoperated on patients who have shown string signs and the grafts look quite small. We have actually taken some sections of them to display those changes. Who is responsible? I don t know. Probably the surgeon, because it s a dilemma, particularly with a big right coronary artery. If you have a 5-mm coronary artery and a 50% to 60% stenosis in a young patient, what do you do? I think that dilemma will stay with us for a long time. The ultimate problem is competitive flow and I think we still are grappling with how to best address that issue in the different age groups. With the right coronary artery, a number of surgeons are saying that perhaps that is not such an important territory; and whether you put a mammary, a radial, a gastroepiploic or vein graft to that site, it probably doesn t matter. I am not sure if that s correct but we will see. How do we overcome the shortcomings? We do use sequential grafting if we can, if the anatomy is appropriate, but again most of these grafts in the angiographic study as I indicated were straight aortocoronary radials and most of those did not have sequential grafting. Antispasmodics? We use nitroglycerin perioperatively for 24 hours intravenously and then amlodipine (which is a once-a-day calcium-channel blocker) for 6 months, based on some empirical observational data that we have had over time. Anticoagulants? We use aspirin 100 mg daily indefinitely. I thank the Society for the opportunity of presenting this paper.