The Radial Artery Versus the Saphenous Vein Graft in Contemporary CABG: A Case-Matched Study

Transcription

1 The Radial Artery Versus the Saphenous Vein Graft in Contemporary CABG: A Case-Matched Study Gideon Cohen, MD, Miguel G. Tamariz, MD, Jeri Y. Sever, Negin Liaghati, BSc, Veena Guru, MD, George T. Christakis, MD, Gopal Bhatnagar, MD, Charles Cutrara, MD, Labib Abouzahr, MD, Bernard S. Goldman, MD, and Stephen E. Fremes, MD Division of Cardiovascular Surgery, Sunnybrook and Women s College Health Sciences Centre, Toronto, Ontario, Canada Background. Although use of the internal thoracic artery has been shown to improve outcomes after coronary artery bypass grafting, the same cannot be said of alternative arterial conduits. To determine the benefit of radial artery (RA) grafting, a case-matched review was undertaken. Methods. Between March 1994 and March 1999, 2,847 patients underwent isolated coronary artery bypass grafting with a left internal thoracic artery graft, plus saphenous vein grafts (SVGs). Of these patients, 478 also received an RA graft (RA group). The RA patients were matched at a ratio of 1:2 with patients receiving only SVGs and a left internal thoracic artery graft (SVG group; n 956) using six prognostic risk factors: age, sex, Canadian Cardiovascular Society class, left ventricular grade, number of diseased vessels, and timing of operation. Target vessels were graded according to quality and graftability and were similar between groups. Outcomes were evaluated by univariate and multivariate analyses. Results. There was a significantly higher prevalence of diabetes, hypertension, and peripheral vascular disease in the RA group (p < 0.05). Although stay in the intensive care unit was shorter in the RA group (RA, 30 2 hours, and SVG, 37 2 hours; p ), total hospital stay was similar between groups. The incidence of perioperative myocardial infarction was higher in the SVG group (SVG, 31 of 956 or 3.2%, and RA, 6 of 478 or 1.3%; p 0.02). Multivariate analysis revealed RA grafting to be protective against early mortality and morbidity (odds ratio 0.58; 95% confidence interval, 0.37 to 0.90; p 0.015) and late mortality and morbidity including late reintervention (risk ratio 0.60; 95% confidence interval, 0.37 to 0.93; p 0.02). Actuarial freedom from events at 36 months postoperatively was greater in the RA group (RA, 95% 2%, and SVG, 86% 4%; p 0.01). Conclusions. Despite a higher prevalence of preoperative comorbidity, patients in the RA group demonstrated improved outcomes after coronary artery bypass grafting. The RA is a viable and beneficial conduit for this operation. (Ann Thorac Surg 2001;71:180 6) 2001 by The Society of Thoracic Surgeons Coronary artery bypass grafting (CABG) is a widespread and established means of treating advanced coronary artery disease. Although the results of contemporary CABG are excellent [1, 2], late benefits are influenced by the fate of the bypass conduits used. Saphenous vein grafts (SVGs), although most commonly employed, are limited by poor long-term patency rates. Up to 15% of SVGs occlude within the first postoperative year [3], and at 10 years postoperatively, only 50% to 60% of SVGs are patent [4, 5]. The clinical consequences of such SVG disease include recurrence of angina, need of percutaneous transluminal coronary angioplasty or reoperation, myocardial infarction (MI), and death [1, 2, 6, 7]. In view of the clear advantages of the internal thoracic Presented at the Thirty-sixth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 31 Feb 2, Address reprint requests to Dr Fremes, Division of Cardiovascular Surgery, Sunnybrook and Women s College Health Sciences Centre, 2075 Bayview Ave, Room H410, Toronto, ON M4N 3M5, Canada; stephen.fremes@swchsc.on.ca. artery (ITA) versus the saphenous vein for left anterior descending coronary artery bypass [6 9], expanded use of arterial conduits has long been advocated. However, for technical reasons as well as a lack of supportive data, multiple arterial grafting for CABG is not routinely performed [10]. Initially used as a bypass conduit in the early 1970s [11], the radial artery (RA) was reintroduced into clinical practice in 1989 [12]. Owing to various anatomical and practical characteristics, the RA represents a potential conduit that may be both technically and clinically advantageous compared with other arterial grafts. The following study summarizes our experiences using the RA as a conduit in combination with a left ITA (LITA) graft and SVGs in patients undergoing CABG. For com- This article has been selected for the open discussion forum on the STS Web site: by The Society of Thoracic Surgeons /01/$20.00 Published by Elsevier Science Inc PII S (00)

2 Ann Thorac Surg COHEN ET AL 2001;71:180 6 RADIAL ARTERY GRAFTING 181 pleteness, comparison was made with patients receiving a LITA graft and SVGs only. Material and Methods Cohort Construction The Division of Cardiovascular Surgery at Sunnybrook and Women s College Health Sciences Centre has maintained a prospective database of all patients undergoing cardiac surgical procedures since its inception in Data are managed and maintained in dbaseiv databases. Between March 1994 and March 1999, 2,847 patients underwent isolated CABG at our institution. Of these, 478 patients were identified as having received an RA graft in addition to a LITA graft with or without additional SVGs (RA group). These 478 patients constituted the case group and were assigned consecutive identification numbers and a group identification code (R). A printout of information was generated on these patients: identification number in the study, date of operation, age, sex, Canadian Cardiovascular Society angina class, left ventricular grade, number of diseased coronary arteries, and timing of operation (elective vs urgent). Four of the six surgeons accounted for 89% of the patients in the RA group, and 95% of the entire surgical population. There were 2,369 patients in the CABG database who received a single LITA graft plus SVGs but no additional arterial grafts (SVG group). During the matching procedure, only the matching variables were listed. There was no additional information screened by the analyst (M.G.T.) for any other prognostic, operative, or outcome data. Each individual RA group patient was matched on the basis of age ( 1 year), sex, Canadian Cardiovascular Society angina class, left ventricular grade, number of diseased coronary arteries, and timing of operation to 2 SVG group patients. An attempt was made to match as closely as possible to the date of operation. The matched patients were assigned the same identification number as their RA group counterpart as well as a group identification code (S). There were 956 patients in the SVG group (controls). The 1:2 matching ratio (RA:SVG) was successful for all 478 patients. Patients who received arterial grafts other than the LITA or RA were excluded. To ensure minimal variability, a subgroup analysis was undertaken whereby target coronary arteries were graded according to quality and graftability in 50 randomly chosen patients from each group. A standardized scoring system based on the percentage of stenosis (0 30% to 49%, 1 50% to 69%, 2 70% to 89%, 3 90% to 99%, 4 100%), target coronary artery size ( mm, to 1.99 mm, to 1.49 mm, mm), target coronary artery quality (0 normal, 1 mildly diseased or irregular vessel, 2 moderately diseased or noncritical stenosis, 3 severely diseased or 50% stenosis), target coronary artery calcification (0 none, 1 mildly calcified, 2 severely calcified), and target location (proximity to coronary sinus) was used for this purpose. Scoring was done by a single blinded observer (G.C.) according to the visual assessment of the coronary angiogram. Specified scores were tallied to determine graftability, and a comparison was made between groups. Surgical Methods The Allen test was used to determine the adequacy of collateral circulation to the hand. The RA was not considered if any of the following conditions applied: the Allen test was positive; the patient had an abnormal upper-extremity Doppler study; a RA plaque was noted on ultrasound; or the patient had a history of vasculitis or Raynaud s disease. The nondominant arm was used almost exclusively for RA harvesting, which was performed simultaneously with harvest of all other conduits. The RA was harvested as a pedicle with adjacent veins and surrounding fatty tissue. An atraumatic no-touch technique was used for harvesting. After heparinization, the artery was gently dilated in situ by a slow intraluminal injection of dilute papaverine hydrochloride and verapamil hydrochloride solution (5 ml: papaverine, 60 mg, and verapamil 5 mg, in 16 ml of Ringer s lactate) [13 15]. Ex vivo, the artery was immersed in the same solution. The saphenous vein was harvested using conventional methods followed by immersion in papaverine solution. More often than not, the RA was used to graft a major coronary branch on the lateral or inferior wall of the heart. Single grafts (one distal anastomosis per graft) as opposed to sequential grafts (more than one distal anastomosis per graft) were constructed. Postoperative Management Electrocardiograms were obtained preoperatively, on day 1 postoperatively, and either at the time of discharge or on day 5 postoperatively. Patients received intravenous nitroglycerin (1 to 4 ug kg 1 min 1 ) for the first 24 hours after operation provided the systolic blood pressure was greater than 100 mm Hg [15 17]. Oral nifedipine (Adalat XL, 20 to 30 mg daily) was continued for 6 months thereafter, beginning on the first postoperative day. For patients intolerant of nifedipine, diltiazem hydrochloride or amlodipine was substituted. Sublingual nifedipine (10 mg every 6 hours) was used in the intensive care unit only in instances of persistent hypertension (systolic blood pressure 140 mm Hg) despite adequate doses of intravenous nitroglycerin or sodium nitroprusside. All patients were placed on a regimen of aspirin, 325 mg daily, beginning 6 hours postoperatively and were maintained indefinitely on cholesterol-lowering agents. Statistical Analyses The SAS for PC [18] and BMD/PLR [19] programs were used for statistical analyses. The baseline characteristics and hospital outcomes for the two groups of patients were compared using 2 analysis or Fisher s exact test for categorical data and t tests for continuous variables. Hospital outcomes were evaluated multivariately by stepwise logistic regression

3 182 COHEN ET AL Ann Thorac Surg RADIAL ARTERY GRAFTING 2001;71:180 6 analysis. Odds ratios and their 95% confidence intervals for hospital events were calculated accordingly. The differences between the SVG and RA groups for late survival and other longitudinal outcomes were evaluated by Kaplan-Meier univariate analyses. The Cox proportional hazards procedure was used to adjust the risk ratios for all prognostic variables including the matching variables using a stepwise procedure. Results are reported as the mean the standard deviation in the text and the table and as the mean the standard error in the figures unless not applicable. Statistical significance was defined as a p value of less than Multivariate procedures used an entry level p value of Results Matching of patients receiving an RA graft in addition to a LITA graft and SVGs (RA group; n 478) was accomplished at a ratio of 1:2 with patients receiving only a LITA graft and SVGs (SVG group; n 956). Preoperative variables are summarized in Table 1. Mean age was years in the SVG group and years in the RA group. Sixteen percent of the patients in both Table 1. Preoperative Clinical Characteristics by Group a,b Characteristic SVG Group (n 956) RA Group (n 478) p Value Age (y) NS Age 70 years 146 (15.3) 64 (13.4) NS Female sex 152 (15.9) 76 (15.9) NS Urgent operation 286 (29.9) 143 (29.9) NS CCS NS CCS (40.1) 190 (39.7) NS CHF 31 (3.2) 13 (2.7) NS Diabetes mellitus 238 (24.9) 160 (33.5) Obesity 526 (55.0) 278 (58.2) NS BSA (m 2 ) NS Hypertension 450 (47.1) 253 (52.9) 0.04 COPD 40 (4.2) 23 (4.8) NS PVD 124 (13.0) 99 (20.7) LV grade NS LV grade 4 8 (0.8) 4 (0.8) NS Redo ACB 20 (2.1) 11 (2.3) NS TVD 301 (31.5) 150 (31.4) NS No. of diseased coronary NS arteries No. of coronary bypasses NS constructed Target score c NS XCL time (min) a Where applicable, data are shown as the mean the standard deviation. b Numbers in parentheses are percentages. c The target score is the mean score of target vessels based on quality and graftability. ACB aortocoronary bypass; BSA body surface area; CCS Canadian Cardiovascular Society angina class; CHF congestive heart failure; COPD chronic obstructive pulmonary disease; LV left ventricular; NS not significant; PVD peripheral vascular disease; RA radial artery; SVG saphenous vein graft; TVD triple-vessel disease; XCL cross-clamp. Fig 1. Perioperative morbidity and mortality by group: There was a significantly higher incidence of perioperative myocardial infarction (MI) and overall morbidity and mortality in the saphenous vein graft group (SVG). (CVA cerebrovascular accident; IABP intraaortic balloon pump; LOS low-output syndrome; RA radial artery group.) groups were female. The frequency of urgent operations was 29.9% in each group. Seventy-six percent of the patients in both groups were in Canadian Cardiovascular Society angina class 3 or 4. There was a significantly higher prevalence of diabetes mellitus (SVG, 238 of 956 or 25%, and RA, 160 of 478 or 34%; p 0.001), peripheral vascular disease (SVG, 124 of 956 or 13%, and RA, 99 of 478 or 21%; p 0.001), and preoperative hypertension (SVG, 450 of 956 or 47%, and RA, 253 of 478 or 53%; p 0.036) in the RA group. Left ventricular grade was scored as grade 1 or 2 in 84% of patients (grade 1 ejection fraction 0.60, grade 2 ejection fraction of 0.40 to 0.60, grade 3 ejection fraction of 0.20 to 0.39, grade 4 ejection fraction 0.20). The mean number of diseased vessels was (range, one to three vessels) in both groups. The mean number of coronary bypass grafts was in the SVG group (range, one to six grafts) and in the RA group (range, two to five grafts) ( p 0.52). The quality and the graftability of target vessels were similar in the two groups (mean scores: SVG, ; and RA, ; p 0.16). Postoperative stay in the intensive care unit was shorter in the RA group (SVG, 37 2 hours, and RA, 30 2 hours; p ), although total hospital stay was similar between groups (SVG, 8 6 days, and RA, 8 5 days; p 0.32). Perioperative mortality was identical in the two groups (SVG, 10 of 956 or 1%, and RA, 5 of 478 or 1%; p 1.00). The incidence of perioperative MI was elevated in the SVG group (SVG, 31 of 956 or 3%, and RA, 6 of 478 or 1%; p 0.02). The incidence of perioperative low-output syndrome (the requirement of inotropic or intraaortic balloon pump support for longer than 30 minutes to sustain a cardiac index greater than 2.2 L min 1 m 2 and a systolic blood pressure greater than 90 mm Hg), however, was not significantly different between groups (SVG, 70 of 956 or 7%, and RA, 25 of 478 or 5%; p 0.133) (Fig 1). Logistic regression analysis revealed RA grafting to be protective against early mortality or morbidity (MI, low-output syndrome, intraaortic balloon pump support,

4 Ann Thorac Surg COHEN ET AL 2001;71:180 6 RADIAL ARTERY GRAFTING 183 Fig 2. Predictors of early mortality and morbidity by logistic regression: Multivariate analysis revealed radial artery grafting and surgeon to be protective against perioperative morbidity (myocardial infarction, low-output syndrome, intraaortic balloon pump requirement, cerebrovascular accident). (CCS 4 Canadian Cardiovascular Society angina class 4; C.I. confidence interval; LV 3-4 left ventricular grade 3 or 4; TVD triple-vessel disease.) Fig 4. Actuarial survival by group: Actuarial survival at 36 months was not significantly different between the radial artery (RA) group and the saphenous vein graft group (SVG). Results are shown as the mean the standard error. stroke) (odds ratio 0.58; 95% confidence interval, 0.37 to 0.90; p 0.015) (Fig 2). Similarly, Cox proportional hazards analysis of hospital survivors demonstrated RA grafting to be protective against late mortality or morbidity (MI, coronary angioplasty, reoperation, readmission for cardiac-related cause) (risk ratio 0.60; 95% CI, 0.37 to 0.93; p 0.02) (Fig 3). Although actuarial survival was not significantly different between groups (SVG, 92% 8%, and RA, 96% 2%; p 0.64), patients in the RA group demonstrated greater actuarial freedom from events including death, MI, coronary angioplasty, reoperation, and cardiacrelated readmission at 36 months postoperatively (SVG, 86% 4%, and RA, 95% 2%; p 0.01) (Figs 4, 5). Comment The RA has long been advocated as a potential graft for CABG. Inherent anatomical and practical features render this conduit technically advantageous compared with alternative arterial grafts. The average length of the graft ( 20 cm) makes it suitable for reaching all myocardial territories [16]. The diameter of the lumen is larger than that of other arterial grafts and corresponds well to most recipient coronary vessels [16]. The thick muscular wall is appropriate for both aortic and coronary anastomoses [16, 17], which are technically less demanding to construct than anastomoses fashioned using the ITAs, right gastroepiploic artery, or inferior epigastric artery. Contraindications to the use of other arterial conduits including advanced age, obesity, diabetes mellitus, obstructive pulmonary disease, and prior laparotomy do not generally apply to the RA. Moreover, relative to saphenous veins, RAs can be harvested without interfering with ambulation, and wound infection is less common. From a practical perspective, the RA can be harvested concurrently with other conduits, and the dissection can usually be accomplished by an assistant, thus preventing any Fig 3. Predictors of late events by Cox regression: Multivariate analysis of hospital survivors revealed radial artery grafting to be protective against late events including late death, myocardial infarction, percutaneous transluminal coronary angioplasty, reoperation, and cardiac-related readmission. (CCS 4 Canadian Cardiovascular Society angina class 4; C.I. confidence interval; PVD peripheral vascular disease.) Fig 5. Actuarial freedom from events by group: Actuarial freedom from death or nonfatal late cardiac-related events (myocardial infarction, percutaneous transluminal coronary angioplasty, reoperation, cardiac-related readmission) at 36 months was significantly greater in the radial artery group (RA) than in the saphenous vein graft group (SVG). Results are shown as the mean the standard error.

5 184 COHEN ET AL Ann Thorac Surg RADIAL ARTERY GRAFTING 2001;71:180 6 undue prolongation of operative time. Finally, the versatility of the RA makes it well suited as a second arterial conduit. Introduced into clinical practice in the early 1970s [11], the RA graft was quickly abandoned because of high occlusion rates compared with contemporary saphenous vein or ITA grafts [20 22]. Carpentier [20], the original user of the RA graft, believed that such high occlusion rates were largely related to graft spasm. This concept was supported when follow-up angiography of patients in whom the RA graft had been thought to be totally or partially occluded revealed patent, angiographically normal RA grafts. In 1989, Carpentier s group reintroduced the RA graft for CABG and reported their experiences in 1992 [12]. Among the 104 patients receiving an RA graft, there was one death (details not provided) and two perioperative MIs, neither of which was attributable to occlusion or spasm of the RA graft. Modifications in technique included harvesting the graft as a pedicle, intraluminal dilation with papaverine (as opposed to mechanical dilation), and postoperative administration of calcium-channel blockers to limit graft spasm. The early patency rate of the first 56 grafts studied within 2 weeks of operation was 100%. Late angiographic follow-up was obtained in 27 patients (31 grafts) at a mean of 9 months postoperatively. These results were better than those observed with SVGs or free ITA grafts. Twenty-nine grafts were patent, and none of them had any degree of narrowing. A later series by Calafiore and coworkers [23] involved 163 RA grafts. At an average of 3.6 months postoperatively, 75 (98.7%) of 76 grafts were patent, and 33 (94.3%) of 35 grafts studied 13 months or more postoperatively (mean time, 21.1 months) were patent. In our series, despite a higher prevalence of preoperative comorbidity, the RA patients demonstrated a lower incidence of perioperative MI. Although the incidence of postoperative low-output syndrome was also slightly lower in the RA group, this difference did not reach significance. Nonetheless, such morbidity did translate into a significantly longer postoperative stay in the intensive care unit for the SVG group. These differences may be attributable to the superior early patency of RA grafts in comparison to SVGs. We have collated all available English-language publications and abstracted relevant data items evaluating the RA (unpublished observations). Cumulative totals suggest that an angiographic patency rate of approximately 97.5% (309 of 317 grafts) can be expected early ( 3 months) postoperatively and 93.0% (186 of 200 grafts) at 1 year. By comparison, in studies by Acar [12], Chen [24], and Affonso da Costa [25], and their associates, the early ( 3 months) patency rate of SVGs was reported as 88.9%, 92.3%, and 92.8%, respectively. Further, in a study by Manasse and colleagues [26], only 35 (76.1%) of 46 SVGs were patent at 1 year postoperatively. In addition to the observed protective nature of RA grafting, one cannot overlook the role of surgeon-related variables to early outcome. The effect of the individual surgeon on the conclusions regarding the beneficial effects of RA grafting were analyzed, as the frequency of RA grafting differed significantly between surgeons. In stratified analyses, RA grafting was protective against early mortality and morbidity after controlling for the variable individual surgeon. In multivariate analyses for early or late events, RA grafting remained protective with minimal change in the odds or risk ratios when variables for individual surgeon were entered into the risk equations, findings suggesting that the RA effects were not a proxy for individual surgeon. Although certain surgeons used the RA more often than others, a consistent relationship could be demonstrated between RA grafting and patient outcome, thus indicating that the results were generalizable, at least within our institution. Our findings at follow-up are not surprising in view of the published long-term patency rates of the RA in comparison to SVGs. Two groups [27, 28] have published RA graft patency results at 5 years postoperatively. In these studies, 111 (88.1%) of 126 RA grafts were demonstrated to be patent at 5 years compared with only 74.1% of vein grafts. In our series, although actuarial survival at 60 months was not significantly different between groups, multivariate analysis revealed RA grafting to be protective against late MI, late readmission, and late reintervention. There was only one late death in the RA group, and it involved a patient with end-stage heart failure. Postmortem examination revealed all grafts to be patent. Of the 20 late readmissions in the RA group, only three were related to recurrence of angina. All 3 patients underwent diagnostic angiography. In 1 patient, the RA graft to the first obtuse marginal artery was found to be occluded. The patient was treated medically. In another patient, 95% stenosis of an RA graft to the right coronary artery was identified, which was treated by angioplasty and stenting of the native coronary artery. In the third patient, occlusion of an SVG to the right coronary artery was identified and was treated by angioplasty of the SVG. Two RA harvest site complications were documented. In 1 patient, a 50-year-old man, a postoperative radial nerve palsy with weakness of extension at the wrist and fingers developed. Although improvement with time has been noted, a return to baseline function had not been achieved up to 5 years postoperatively. The second patient, a 67-year-old man, required readmission to the hospital for management of superficial cellulitis at the site of RA harvesting. The treatment, consisting of intravenous antibiotics, was successful, and no long-term sequelae were experienced. The primary limitation of the study is its retrospective nature along with the potential for bias in the selection of patients for RA grafting. The study design did attempt to control for bias by using matched controls and by employing a semiquantitative scoring method for target vessels in a subset of patients. To alleviate such limitations, our group [29] has organized a multicenter randomized clinical trial comparing the 8- to 12-month angiographic patency of the RA graft versus the SVG for bypass of targets other than the left anterior descending coronary artery. This study should provide an unbiased

6 Ann Thorac Surg COHEN ET AL 2001;71:180 6 RADIAL ARTERY GRAFTING 185 estimate of the relative patency of RA grafts while controlling for patient variables and recipient vessel variables. Despite a higher prevalence of preoperative comorbidity, patients receiving an RA graft during CABG demonstrated improved early and late outcomes. In summary, our findings, in combination with previously published reports, suggest that RA grafting (mainly to targets other than the left anterior descending coronary artery) is better than saphenous vein grafting for CABG and may be comparable to right ITA grafting. Supported in part by grant MI from the Medical Research Council of Canada. References 1. Kirklin JW, Naftel DC, Blackstone EH, Pohost GM. Summary of a consensus concerning death and ischemic events after coronary artery bypass grafting. Circulation 1989;79(Suppl 1): Kirklin JW, Akins CW, Blackstone EH, et al. ACC/AHA Task Force report. Guidelines and indications for coronary artery bypass graft surgery. J Am Coll Cardiol 1991;17: Fremes SE, Levinton C, Naylor CD, et al. Optimal antithrombotic therapy following aortocoronary bypass: a metaanalysis. Eur J Cardiothorac Surg 1993;7: Bourassa MG, Fisher LD, Campeau L, Gillespie MJ, McConney M, Lesperance J. Long-term fate of bypass grafts: the Coronary Artery Surgery Study (CASS) and Montreal Heart Institute experiences. Circulation 1985;72(6 pt 2):V Kouchoukos NT, Karp RB, Oberman A, Russell RO Jr, Alison HW, Holt JH Jr. Long-term patency of saphenous veins for coronary bypass grafting [Abstract]. Circulation 1977; 56(Suppl 3): 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: Cameron A, Davis KB, Green G, Schaff HV. Coronary bypass surgery with internal-thoracic-artery grafts effects on survival over a 15-year period. N Engl J Med 1996;334: Grondin CM, Campeau L, Lesperance J, Enjalbert M, Bourassa MG. 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(3 Pt 2):I Myers WO, Blackstone EH, Davis K, Foster ED, Kaiser GC. CASS Registry long term surgical survival. Coronary Artery Surgery Study. J Am Coll Cardiol 1999;33: Grover FL, Johnson RR, Marshall G, Hammermeister KE, Department of Veterans Affairs Cardiac Surgeon. Impact of mammary grafts on coronary bypass operative mortality and morbidity. Ann Thorac Surg 1994;57: Carpentier A, Guermonprez JL, Deloche A, Frechette C, DuBost C. The aorta-to-coronary radial artery bypass graft: a technique avoiding pathological changes in grafts. Ann Thorac Surg 1973;16: Acar C, Jebara VA, Portoghese M, et al. Revival of the radial artery for coronary artery bypass grafting. Ann Thorac Surg 1992;54: Fremes SE, Christakis GT, Del Rizzo DF, Musiani A, Mallidi H, Goldman BS. The technique of radial artery bypass grafting and early clinical results. J Cardiac Surg 1995;10: He G-W, Yang C-Q. Use of verapamil and nitroglycerin solution in preparation of radial artery for coronary grafting. Ann Thorac Surg 1996;61: Cable DG, Caccitolo JA, Pearson PJ, et al. New approaches to prevention and treatment of radial artery graft vasospasm. Circulation 1998;98(19 Suppl):II Grooters RK, Nishida H. The radial artery. In: Grooters RK, Nishida H, eds. Alternative bypass conduits and methods for surgical coronary revascularization. Mt. Kisco, NY: Futura, 1994; Van Son JAM, Smedts F, Vincent JG, van Lier HJJ, Kubat K. Comparative anatomic studies of various arterial conduits for myocardial revascularization. J Thorac Cardiovasc Surg 1990;99: SAS language guide for personal computers. Cary, NC: SAS Institute Inc, Dixon WJ, ed. BMDP statistical software manual. Berkeley, CA: University of California Press, Carpentier A. Discussion of Geha AS, Krone RJ, McCormick JR, Baue AE. Selection of coronary bypass: anatomic, physiological, and angiographic considerations of vein and mammary artery grafts. J Thorac Cardiovasc Surg 1975;70: Curtis JJ, Stoney WS, Alford WC Jr, Burrus GR, Thomas CS Jr. Intimal hyperplasia: a cause of radial artery aortocoronary bypass graft failure. Ann Thorac Surg 1975;20: Fisk RL, Brooks CH, Callaghan JC, Dvorkin J. Experience with the radial artery graft for coronary artery bypass. Ann Thorac Surg 1976;21: Calafiore AM, Di Giammarco G, Teodori G, et al. Radial artery and inferior epigastric artery in composite grafts: improved midterm angiographic results. Ann Thorac Surg 1995;60: Chen AH, Tatsuya N, Brodman RF, et al. Early postoperative angiographic assessment of radial artery grafts used for coronary artery bypass grafting. J Thorac Cardiovasc Surg 1996;111: da Costa FD, da Costa IA, Poffo R, et al. Myocardial revascularization with the radial artery: a clinical and angiographic study. Ann Thorac Surg 1996;62: Manasse E, Sperti G, Suma H, et al. Use of the radial artery for myocardial revascularization. Ann Thorac Surg 1996;62: Possati G, Guadino M, Alessandrini F, et al. Midterm clinical and angiographic results of radial artery grafts used for myocardial revascularization. J Thorac Cardiovasc Surg 1998;116: Acar C, Ramsheyi A, Pagny JY, et al. The radial artery for coronary artery bypass grafting: clinical and angiographic results at 5 years. J Thorac Cardiovasc Surg 1998;116: Radial Artery Patency Study Investigators. Multicentre Radial Artery Patency Study (RAPS): study design. Control Clin Trials 2000;21: DISCUSSION DR BRIAN F. BUXTON (Heidelberg, Victoria, Australia): Doctor Cohen, I enjoyed your paper very much. My colleagues and I are very aware of the work that you and your group, including Dr Fremes, are doing in this area. With any retrospective casematched study, one of the problems is to know which variables to put into the matching process. One of the main concerns when using arterial grafts is the degree of stenosis in the native vessel. For instance, if there are large numbers of native vessels with a low-grade stenosis in the radial artery group, you might expect many of those grafts to fail. Have you found a way of

7 186 COHEN ET AL Ann Thorac Surg RADIAL ARTERY GRAFTING 2001;71:180 6 analyzing, even retrospectively, the degree of stenosis in the grafted vessels? DR COHEN: Thank you for your question. The degree of inherent stenosis within the bypassed coronary vessels was, in fact, analyzed, but this was not included in the scoring system applied to each target vessel. Indeed, there was a tendency towards the use of radial artery grafts to bypass only those vessels with high-grade stenoses. This has been the policy at our institution, and it continues to serve as a prerequisite for radial artery grafting in our ongoing prospective, randomized trial. At this point, I cannot comment on the fate of the radial artery when applied to vessels with low-grade stenoses because of the relatively small numbers of patients in this category. After having reviewed selected postoperative angiograms, I believe that excessive competitive flow within the native coronary artery may contribute to the characteristic string-sign effect often identified in failed radial artery bypass grafts. However, I do not have any objective data with which to confirm or refute this belief. DR BUXTON: My second query pertains to the long-term outcome of the study. You had 3-year clinical end points. Do you have any late graft patency studies on these patients? DR COHEN: Yes, we do. As I mentioned, we currently are undertaking a multicenter prospective, randomized trial in which the radial artery is being compared with the saphenous vein for coronary revascularization, with each patient acting as his or her own internal control. The primary end point of this study is angiographic patency. We hope to publicize the results of this trial in the near future. DR JOHN PUSKAS (Atlanta, GA): Dr Cohen, would you please clarify the following points: the technique used to harvest the radial arteries, the pharmacologic maneuvers used to dilate the artery before implantation, and the therapy you apply in the postoperative period for patients who have received radial arteries. I enjoyed your presentation. DR COHEN: Thank you for your questions. We use a notouch technique to harvest the radial artery. The saphenous vein is harvested using conventional methods. After mobilization of its distal aspect, the radial artery is gently dilated using a solution comprising 60 mg of papaverine hydrochloride and 5 mg of verapamil hydrochloride. The radial artery is then immersed in the same solution until final use. In situations where radial anastomosis may be delayed, the vessel is attached to a manifold off the main arterial line for continuous infusion off the bypass circuit. Postoperatively, if tolerated, all patients receive intravenous nitroglycerin at a dosage of 1 to 4 ug kg 1 min 1 during their stay in the intensive care unit. On the first postoperative day, patients are routinely placed on a calcium-channel blocker. At our institution, we use nifedipine, 20 to 30 mg orally once a day. This regimen is maintained for a period of 6 months postoperatively.