Atrial fibrillation (AF) is a frequent complication of

Predictors of Atrial Fibrillation After Conventional and Beating Heart Coronary Surgery A Prospective, Randomized Study Raimondo Ascione, MD; Massimo Caputo, MD; Giliola Calori, MD; Clinton T. Lloyd, FRCS; Malcom J. Underwood, FRCS; Gianni D. Angelini, FRCS Background Atrial fibrillation (AF) increases the morbidity of CABG. The pathophysiology is uncertain, and its prevention remains suboptimal. This prospective, randomized study was designed to define the role of cardiopulmonary bypass (CPB) and cardioplegic arrest in the pathogenesis of this complication. Methods and Results Two hundred patients were prospectively randomized to (1) on-pump conventional surgery [(100 patients, 79 men, mean age 63 (40 to 77) years)] with normothermic CPB and cardioplegic arrest of the heart or (2) off-pump surgery [(100 patients, 82 men, mean age 63 (38 to 86) years)] on the beating heart. Heart rate and rhythm were continuously monitored with an automated arrhythmia detector during the first 72 hours after surgery. Thereafter, routine clinical observation was performed and continuous monitoring restarted in the case of arrhythmia. The association of perioperative factors with AF was investigated by univariate analysis. Significant variables were then included into a stepwise logistic regression model to ascertain their independent influence on the occurrence of AF. There were no significant baseline differences between groups. Thirty-nine patients in the on-pump group and 8 patients in the off-pump group had postoperative sustained AF (P 0.001). Univariate analysis showed that CPB inclusive of cardioplegic arrest, postoperative inotropic support, intubation time, chest infection, and hospital length of stay were predictors of AF (all P 0.05). However, stepwise multivariate regression analysis identified CPB inclusive of cardioplegic arrest as the only independent predictor of postoperative AF (OR 7.4; CI 3.4 to 17.9). Conclusions CPB inclusive of cardioplegic arrest is the main independent predictor of postoperative AF in patients undergoing coronary revascularization. (Circulation. 2000;102:1530-1535.) Key Words: prevention tachyarrhythmias fibrillation coronary disease cardiopulmonary bypass Atrial fibrillation (AF) is a frequent complication of CABG. 1,2 Its incidence varies depending on the definition used, the mode of monitoring, and the clinical profile of patients. 3 Although it is not a life-threatening event, it may lead to hemodynamic compromise, thromboembolic events, anxiety, and increased costs. 1 3 The underlying cause of AF has been related to a variety of preoperative and postoperative factors. 4 7 Strategies directed toward management and reduction of postoperative AF have focused mainly on antiarrhythmic drugs. 1,2,8,9 Myocardial ischemia and inadequate cardioplegic protection of the atria have been reported to increase the incidence of postoperative AF. 3,10 Myocardial revascularization on the beating heart does not require atrial cannulation, cardiopulmonary bypass (CPB), and cardioplegic arrest, 11 15 and it has been suggested to be associated with a reduction in the incidence of postoperative AF. 12,14 The present prospective, randomized study investigated the incidence of AF in similar cohorts of patients undergoing CABG with or without CPB and cardioplegic arrest. Further analysis included variables considered by other authors to have a possible association with AF. Methods Patient Selection Over a 19-month period (March 1997 to August 1998), 538 patients underwent first-time CABG under the supervision of a single consultant. On the basis of their eligibility for off-pump surgery, 200 patients were individually prospectively randomized for myocardial revascularization with either (1) on-pump conventional surgery (100 patients, 79 men, mean age 63 [40 to 77] years) with normothermic CPB and cardioplegic arrest of the heart or (2) off-pump surgery (100 patients, 82 men, mean age 63 [38 to 86] years) on the beating heart. The randomization sequence was obtained by card allocation and strictly respected. Exclusion from the randomization study was based on criteria that included history of supraventricular arrhyth- Received February 9, 2000; revision received April 26, 2000; accepted May 2, 2000. From Bristol Heart Institute, Bristol Royal Infirmary, Bristol, UK (R.A., M.C., C.T.L., M.J.U., G.D.A.), and Epidemiology Unit, San Raffaele Hospital, IRCCS, Milan, Italy (G.C.). Correspondence to Prof Gianni D. Angelini, Bristol Heart Institute, Bristol Royal Infirmary, Bristol BS2 8HW, UK. E-mail N.Holloway-Dee@ bristol.ac.uk 2000 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org 1530

Ascione et al Perioperative Predictors of Atrial Fibrillation 1531 mia, left ventricular ejection fraction of 30%, recent myocardial infarction (MI) ( 1 month), repeat operation, renal and respiratory impairment, previous stroke or transient ischemic attack, and coagulopathy. Patients with coronary disease involving branches of the circumflex artery distal to the first obtuse marginal branch and posterior branches originating from the left system were also excluded from the study because these were believed to be technically difficult for adequate revascularization at the beginning of our experience with off-pump surgery. The study was approved by the United Bristol Healthcare Trust Ethics Committee, and all patients gave informed consent. Management of Preoperative and Postoperative Medications Preoperative medications including -blockers, diuretics, antihypertensives, and calcium channel blockers were routinely omitted on the day of surgery. ACE inhibitors were withdrawn on the evening before the operation. On the first postoperative day, in accordance with the intensive care unit protocol (if heart rate 55 bpm, systolic blood pressure 110 mm Hg), -blockers and antihypertensive drugs were restarted. Anesthetic Technique In both groups, anesthetic technique consisted of propofol infusion at 3mg kg 1 h 1 combined with remifentanil infusion at 0.5 to 1 g kg 1 min 1. Neuromuscular blockade was achieved by 0.1 to 0.15 mg/kg pancuronium bromide or vecuronium and the lungs ventilated to normocapnia with air and oxygen (45% to 50%). Heparin was given at a dose of 300 IU/kg to achieve a target activated clotting time (ACT) of 480 seconds before commencement of CPB in the on-pump group. An additional 3000 IU of heparin was administered if required. In the off-pump group, heparin (100 IU/kg) was administered before the start of the first anastomosis to achieve an ACT of 250 to 350 seconds. On completion of all anastomoses, protamine was given to reverse the effect of heparin and return the ACT to preoperative levels. Operative Technique CPB was instituted with the use of ascending aortic cannulation and 2-stage venous cannulation of the right atrium. A standard circuit was used: a Bard tubing set, which included a 40- m filter (Sorin Biomedica), and a hollow-fiber membrane oxygenator. Nonpulsatile flow was used, and flow rates throughout bypass were 2.4 L/m 2 per minute. Systemic temperature was kept between 34 and 36 C. Myocardial protection was achieved by using intermittent anterograde hyperkalemic warm blood cardioplegia. 16 The method of exposure and stabilization to perform the anastomosis consisted of the technique previously described by our group 17 with or without the use of the Cardiothoracic System (CTS) retractor and stabilizer. The target vessel was then exposed and snared above the anastomotic site with a 4-0 prolene suture with a soft plastic snugger to prevent coronary injury. The coronary artery was then opened, and the anastomosis was performed. Visualization was enhanced by the use of a surgical blower-humidifier (Research Medical Inc). Postoperative Management At the end of surgery, patients were transferred to the intensive care unit (ICU). The lungs were ventilated with 60% oxygen with volume-controlled ventilation and a tidal volume of 10 ml/kg with 5cmH 2 O of positive end-expiratory pressure. Adjustments in FIO 2 and respiratory rate were made according to routine blood gas analysis to maintain PaO 2 between 80 and 100 mm Hg and PaCO 2 between 35 and 40 mm Hg. Forced air warming was used until a stable nasopharyngeal temperature of 37 C was reached. Patients were extubated as soon as they met the following criteria: hemodynamic stability, no excessive bleeding ( 80 ml/h), normothermia, TABLE 1. Baseline Characteristics (n 100) (n 100) P * Age, y 63(40 77) 63 (38 86) 0.8 Sex, M/F 79/21 82/18 0.6 Congestive heart failure 6 5 0.7 Previous MI 30 41 0.1 Diabetes mellitus 13 19 0.2 Hypertension 48 62 0.05 Hypercholesterolemia 70 77 0.3 Smoking 14 10 0.3 Priority Routine admission 63 65 0.7 Unstable in hospital 37 35 0.7 Angina class 0.07 I 12 7 II 40 38 III 34 41 IV 14 14 Ejection fraction Good ( 50%) 79 81 0.7 Moderate (30 50%) 21 19 0.7 -Blocker drugs 39 32 0.3 Diuretics 7 13 0.2 Nitrates 82 75 0.2 ACE inhibitors 33 33 1.0 Parsonnet score 4 (0 15) 5 (0 34) 0.3 Data are presented as median with minimum-maximal values or number of patients (in parentheses). and consciousness with pain control. Fluid management after surgery consisted of 5% dextrose infused at 1 ml kg 1 h 1, with additional colloid solution or blood to maintain normovolemia and hematocrit 24%. Potassium and magnesium deficiency was promptly treated as necessary to maintain electrolyte balance within the normal range. Monitoring and Definitions Heart rate and rhythm were continuously monitored and displayed on a screen with an automated arrhythmia detector (Solar 8000 Patient Monitor, Marquette Medical Systems) during the first 72 hours after surgery. Automatic printing of the ECG was related to the inserted alarm levels and included heart rate 90 bpm and the presence of 6 consecutive normal R-R intervals varying by 100 ms. Twelve-lead ECG recordings were performed before surgery, 2 hours after surgery, and then daily thereafter until hospital discharge. After the first 72 hours, trained nurses performed clinical observations every 4 hours. An ECG was recorded on the basis of any clinical suspicion of arrhythmia. In the case of documented arrhythmia, continuous ECG monitoring was restarted. Each episode of arrhythmias was printed out and interpreted by an independent intensive care physician. AF, atrial flutter, and atrial tachycardia were defined according to Kalman et al. 2 AF was defined as nonsustained if lasting between 10 beats and 10 minutes and sustained if persisting for 10 minutes. 2 For the purpose of the analysis, only a first event was recorded, provided its duration satisfied prespecified criteria. Indications for temporary pacing were made by an independent intensive care physician, according to the ICU protocol, and included symptomatic bradycardia, unresponsive to treatment by drugs, acute

1532 Circulation September 26, 2000 conduction disturbances including second- or third-degree atrioventricular (AV) block, and bifascicular or trifascicular block. Indications for permanent pacemaker were made by an independent cardiologist and included second-degree (Mobitz type II) and thirddegree AV block and long-lasting symptomatic bradycardia. Intraoperative and postoperative data, including complications and adverse events, were recorded. Clinical diagnostic criteria for perioperative MI were new Q waves of 0.04 ms and/or a reduction in R waves 25% in 2 leads. Chest infection was defined as the presence of purulent sputum associated with fever and requiring antibiotic therapy according to positive sputum culture. Statistical Analysis Arrhythmias such as atrial flutter, atrial tachycardia, or others were not considered in the same group of AF because their mechanism differs. 2 Only sustained episodes of AF were included in the statistical analysis. The association of preoperative, intraoperative, and postoperative variables with postoperative AF was investigated with univariate analysis. The following factors, reported as predictors of AF by other investigators, were also included in the analysis: age at operation, preoperative use of -blocker, left ventricular ejection fraction, previous MI, diabetes mellitus, surgery of the right coronary artery/posterior descending coronary artery (RCA/PDA), number of grafts, CPB and cardioplegic arrest, intubation time, chest infection, inotropic requirement, and blood loss. Because continuous data did not show a gaussian distribution, they were analyzed by the Mann-Whitney test. Categoric data comparison was made by 2 test. Data are expressed as median with minimum-maximal values or number (percentage) of patients. Factors that turned out to be predictors of AF at univariate analysis were then included in a stepwise logistic multivariate regression model to ascertain their independent role. ORs and 95% CIs were calculated. Results Preoperative and operative characteristics are shown in Tables 1 and 2, respectively. The two groups were similar with respect to age, sex, and severity of coronary disease, diabetes mellitus, angina class, and surgical data, including number of distal anastomoses. There were 2 deaths in the on-pump group, both for multiorgan failure as a consequence of low cardiac output. Four patients in the on-pump group and 1 in the off-pump group had perioperative MI as per predefined criteria. TABLE 3. Postoperative Arrhythmias TABLE 2. Intraoperative Data (n 100) (n 100) P * CPB time, min 65 (23 202) Cross-clamp time, min 40 (13 84) Grafts per patient, n 2 (1 4) 2 (1 4) 0.3 IMA to LAD or diag graft, n 98 98 1.0 SV to diag graft, n 25 22 0.7 RCA/PDA graft, n 62 62 1.0 Proximal Cx graft, n 43 37 0.1 Other arterial grafts, n 22 15 0.1 Data are presented as median with minimum-maximal values or number of patients (in parentheses). IMA indicates internal mammary artery; LAD, left anterior descending coronary artery; diag, diagonal branch of LAD; SV, saphenous vein; and Cx, circumflex artery. This included the use of left and right IMA, radial artery, and gastroepiploic artery to coronary vessels other than LAD. The incidence of perioperative arrhythmias is presented in Table 3. The overall incidence of AF was 28% (56 patients), with 45 patients in the on-pump group and 11 patients in the off-pump group (P 0.001). Of these, 39 in the on-pump group and 8 in the off-pump group had sustained episodes of AF (P 0.001), and the average of their duration was similar between the two groups (Table 3). Atrial flutter occurred in 2 patients in the on-pump group and 1 patient in the off-pump group. Temporary pacing was required in 7 patients in the on-pump group (1 transient second-degree AV block and 6 transient symptomatic bradycardia) over the first 24 hours after surgery. One patient was temporarily paced in the off-pump group because of transient second-degree AV block (P 0.03, on versus off pump). One patient in each group was affected by third-degree AV block requiring a permanent pacemaker. Postoperative clinical data are reported in Table 4. The off-pump group showed reduced inotropic usage, chest infec- (n 100 patients) (n 100 patients) P * Arrhythmia incidence, n 49 14 0.001 Total AF, n 45 11 0.001 Sustained AF, n 39 8 0.001 Unsustained AF, n 6 3 0.3 Duration of sustained AF, h 14 (6 67) 13 (8 71) 0.3 Atrial flutter, n 2 1 0.8 Transient second-degree AV 1 1 1.0 block, n Third-degree AV block, n 1 1 1.0 Permanent pacemaker, n 1 1 1.0 Temporary pacing over first 24 h 7 1 0.03 Data are presented as median with minimum-maximal values or number of patients (in parentheses). * 2 test.

Ascione et al Perioperative Predictors of Atrial Fibrillation 1533 TABLE 4. Postoperative Data (n 100) (n 100) P * Deaths 2 0 0.7 MI 4 1 0.2 Chest infection 22 8 0.006 Transient stroke 2 2 1.0 Inotropic requirement 23 6 0.001 Total blood loss, ml 900 (90 4220) 690 (110 2300) 0.0001 Intubation time, h 8 (3 192) 6 (3 21) 0.0001 ICU length of stay, d 1 (1 8) 1 (0 6) 0.0001 Hospital length of stay, d 7 (5 17) 5 (4 23) 0.0001 Data are presented as median with minimum-maximal values or number of patients (in parentheses). tion, blood loss and transfusion requirement, intubation time, and length of ICU and hospital stay. Results from the univariate analysis for the association of each factor with AF is reported in Table 5. The use of CPB inclusive of cardioplegic arrest, postoperative inotropic requirement, prolonged intubation time, chest infection, and hospital permanence were all predictors of AF. However, stepwise multivariate regression analysis indicated the use of CPB inclusive of cardioplegic arrest as the only independent predictor of postoperative AF (OR 7.4; CI 3.4 to 17.9). Discussion AF is the most common complication occurring after cardiac surgery. 1 3 Despite advances in CPB, cardioplegic arrest, and surgical techniques, its incidence has paradoxically increased in recent years 18 as the result of surgical patients being older and sicker and advances in ECG continuous monitoring technology. 3 It is frequently not well tolerated, and patients may have symptoms including temporary hemodynamic instability, thromboembolic events, and shortness of breath or chest discomfort and has been shown to increase costs and to lengthen hospital stay. 1,3 Many preoperative and postoperative factors have been suggested to increase the incidence of postoperative AF after conventional CABG such as advanced age, hypertension, 3 withdrawal of -blocker drug, 19 RCA stenosis, 5 respiratory complications, 6 and bleeding. 7 Strategies directed toward reduction of postoperative AF have focused on several drugs, given prophylactically, such as -adrenoceptor antagonists, 1,19 calcium antagonists, 9 amiodarone, and propafenone, 10,20 with conflicting results. 1 However, little is known about intraoperative mechanisms through which the incidence of postoperative AF could be reduced. 21 Recently, there has been a renewed interest in performing CABG on the beating heart. 11 14 Several studies have reported improved myocardial and renal protection, 11,22 minimal inflammatory response, 2 24 excellent patency rate of the grafts, 13 and decreased costs. 3,25 Furthermore, a reduction in postoperative AF has been reported in off-pump series, 12,14 although this remains an area of controversy. 15 Buffolo and coworkers 12 reported a significantly lower incidence of arrhythmias in a cohort of 1274 patients undergoing off-pump surgery when compared with patients undergoing CABG on pump. Abreu et al 14 have also reported a trend toward a lower incidence of postoperative AF in patients undergoing CABG without CPB. Conversely, Cohn et al, 15 in a very recent retrospective age-matched study, failed to demonstrate a lower incidence of postoperative AF among patients undergoing minimally invasive direct CABG than among those undergoing conventional CABG. This led the authors to conclude that mechanisms common to both groups must be considered as determinants of postoperative AF. Whereas the above studies, 12,14,15 by comparing conventional CABG with beating coronary surgery, are appropriate in terms of control group, in none of these was a prospective randomization undertaken. To the best of our knowledge, the present study is the first prospective, randomized study to investigate the role of intraoperative factors such as CPB and associated techniques as determinants of postoperative AF in similar cohorts of patients undergoing CABG with or without CPB through a median sternotomy. The main finding of the present study is that CPB inclusive of cardioplegic arrest is the main independent predictor for the development of postoperative AF, the risk being 7 times higher in the on-pump compared with the off-pump group. This could be related to the period of myocardial ischemia or to the inflammatory response after CPB itself, the required atrial cannulation, and the adverse effects of cardioplegia. Inadequate atrial protection has been postulated as the trigger responsible for the development of AF in vulnerable patients. 26,27 Myocardial damage has been reported to occur after cross-clamping of the aorta 28 and cardioplegic arrest, 1,2 with its effects influencing myocardial ischemia and arrhythmias during both reperfusion and the early postoperative periods. 9 In a recent study, we have shown a significantly higher troponin I release in patients undergoing conventional CABG when compared with beating heart revascularization. 11 Sympathetic activation may be important in the pathogenesis of AF after CABG, 2,6 and this underlines the importance of -adrenoceptor blockade as prophylaxis. White and associates 19 found a significant increase in the incidence of AF in patients in whom administration of -adrenoceptor antagonists was ceased when compared with those whose drug treatment was continued after surgery. In the present study, preoperative -blockers were administered until the day of surgery and restarted on the first postoperative day unless in the presence of slow heart rate or hemodynamic compromise. In keeping with the findings of Kalman et al 2 and Aranki et al, 3 the univariate analysis of the present study showed that the need for postoperative inotropic requirement, ventilation time, chest infection, and hospital stay were all predictors of development of AF. However, when these factors were included in a stepwise logistic multivariate regression model, they did not appear to have an independent role. In contrast with previous reports, 1 7 this study failed to demonstrate advanced age as an independent predictor of postoperative AF, although overall a higher percentage of AF was observed in older patients.

1534 Circulation September 26, 2000 TABLE 5. Univariate Analysis for Atrial Fibrillation AF (n 47 patients) Non-AF (n 153 patients) P * Preoperative Age 65 (45 75) 61 (38 86) 0.34 Sex, M/F 36/11 125/28 0.44 Previous MI 14 (30%) 57 (37%) 0.35 Diabetes mellitus 5 (11%) 27 (18%) 0.25 Hypertension 27 (57%) 83 (54%) 0.7 Smoking 5 (11%) 19 (12%) 0.74 Dyslipidemia 31 (66%) 116 (76%) 0.18 -Blockers 21 (45%) 50 (33%) 0.13 Diuretics 3 (6%) 17 (11%) 0.33 Nitrates 40 (85%) 117 (76%) 0.21 ACE inhibitors 11 (23%) 55 (36%) 0.11 Left ventricular function 37 (79%) 121 (79%) 0.96 Angina class 28 (60%) 73 (48%) 0.15 Parsonnet score 5 (0 13) 4 (0 34) 0.84 Intraoperative CPB cardioplegic arrest 39 (83%) 61 (40%) 0.001 No. of grafts 2 (1 4) 2 (1 4) 0.87 Revascularization of RCA/PDA 27 (57%) 97 (63%) 0.46 Aortic cross-clamp time 39 (14 84) 42 (13 84) 0.40 CPB time 64 (26 202) 65 (23 120) 0.82 Postoperative Chest infection 12 (26%) 18 (12%) 0.02 Inotropic support 12 (26%) 17 (11%) 0.01 Blood loss 850 (400 4220) 750 (90 3000) 0.06 Intubation time 8 (3 192) 6 (3 30) 0.0004 ICU stay 1 (1 8) 1 (0 6) 0.10 Hospital stay 7 (5 13) 6 (4 23) 0.003 Data are presented as median with minimum-maximal values (in parentheses) or number of patients with percentage (in parentheses). The 39% incidence of sustained episodes of AF in the on-pump group is in the upper range of that previously reported when similar cardioplegia was used. 29 This relatively high incidence may reflect the use of continuous monitoring during the first 3 postoperative days. The relation between AF and postoperative bleeding reported by others 7 did not quite reach statistical significance in the univariate analysis of the present study (P 0.06). However, the blood loss of the on-pump group in this study is in keeping with other recently published data. 30 A significant percentage of our patients were in-hospital referrals with unstable angina who were still receiving aspirin and lowmolecular heparin. Furthermore, cell-savers, tranexamic acid, or aprotinin were not used in this study. The limitation of the present study is the absence of continuous Holter ECG monitoring 31 after the first 72 hours until discharge from the hospital. Although it is possible that short episodes of asymptomatic AF might have been missed, this is unlikely for sustained episodes, given the fact that patients were assessed every 4 hours. Furthermore, if there was any clinical suspicion of arrhythmia, an ECG was promptly recorded. Conclusions This prospective, randomized study clearly shows CPB inclusive of cardioplegic arrest as the main independent predictor of postoperative AF in patients undergoing CABG. The minimal incidence of AF in the off-pump group satisfies the modern demand of further improvements in perioperative patient treatment, reduction of costs and resources while maintaining quality of care, and patient satisfaction. Acknowledgments The Garfield Weston Trust and the British Heart Foundation supported this work. References 1. Pfisterer ME, Kloter-Weber UCD, Huber M, et al. Prevention of supraventricular tachyarrhythmias after open heart operation by low-dose sotalol: a prospective, double-blind, randomised, placebo-controlled study. Ann Thorac Surg. 1997;64:1113 1119.

Ascione et al Perioperative Predictors of Atrial Fibrillation 1535 2. Kalman JM, Munawar M, Howes LG, et al. Atrial fibrillation after coronary artery bypass grafting is associated with sympathetic activation. Ann Thorac Surg. 1995;60:1709 1715. 3. Aranki SF, Shaw DP, Adams DH, et al. Predictors of atrial fibrillation after coronary artery surgery: current trends and impact on hospital resources. Circulation. 1996;94:390 397. 4. Crosby LH, Pifalo WB, Woll KR, et al. Risk factors for atrial fibrillation after coronary artery bypass grafting. Am J Cardiol. 1990;66:1520 1522. 5. Mendes LA, Connelly GP, McKenney PA, et al. Right coronary artery stenosis: an independent predictor of atrial fibrillation after coronary artery bypass surgery. J Am Coll Cardiol. 1995;25:198 202. 6. Almassi GH, Schowalter T, Nicolosi AC, et al. Atrial fibrillation after cardiac surgery: a major morbid event? Ann Surg. 1997;226:501 511. 7. Moulton MJ, Creswell LL, Machery ME, et al. Reexploration for bleeding is a risk factor for adverse outcomes after cardiac operations. J Thorac Cardiovasc Surg. 1996;111:1037 1046. 8. Daoud EG, Strickberger SA, Man KC, et al. Preoperative amiodarone as prophylaxis against atrial fibrillation after heart surgery. N Engl J Med. 1997;337:1785 1791. 9. Seitelberger R, Hannes W, Gleichauf M, et al. Effects of diltiazem on perioperative ischemia, arrhythmias, and myocardial function in patients undergoing elective coronary bypass grafting. J Thorac Cardiovasc Surg. 1994;107:811 821. 10. Rousou JA, Meeran MK, Engelman RH, et al. Does the type of venous drainage or cardioplegia affect postoperative conduction and atrial arrhythmias? Circulation. 1985;72(suppl II):II-259-II-263. 11. Ascione R, Lloyd CT, Gomes WJ, et al. Beating versus arrested heart revascularization: evaluation of myocardial function in a prospective randomised study. Eur J Cardiothorac Surg. 1999;15:685 690. 12. Buffolo E, de Andrade CS, Branco JN, et al. Coronary artery bypass grafting without cardiopulmonary bypass. Ann Thorac Surg. 1996;61: 63 66. 13. Poirier NC, Carrier M, Lesperance J, et al. Quantitative angiographic assessment of coronary anastomoses performed without cardiopulmonary bypass. J Thorac Cardiovasc Surg. 1999;117:292 297. 14. Abreu JE, Reilly J, Salzano RP, et al. Comparison of frequencies of atrial fibrillation after coronary artery bypass grafting with and without the use of cardiopulmonary bypass. Am J Cardiol. 1999;83:775 776. 15. Cohn WE, Sirois CA, Johnson RG. Atrial fibrillation after minimally invasive coronary artery bypass grafting: a retrospective, matched study. J Thorac Cardiovasc Surg. 1999;117:298 301. 16. Caputo M, Dihmis W, Bryan AJ, et al. The effect of magnesium on myocardial metabolism during ischemia and reperfusion, in patients undergoing coronary artery surgery using intermittent antegrade warm blood cardioplegia. Eur J Cardiothorac Surg. 1998;14:596 601. 17. Lucchetti V, Angelini GD. An inexpensive method of heart stabilisation during coronary artery operation without cardiopulmonary bypass. Ann Thorac Surg. 1998;65:1477 1478. 18. Creswell LL, Schuessler RB, Rosenbloom M, et al. Hazards of postoperative atrial arrhythmias. Ann Thorac Surg. 1993;56:539 549. 19. White HD, Antman EM, Glynn MA, et al. Efficacy and safety of timolol for prevention of supraventricular tachyarrhythmias after coronary artery bypass surgery. Circulation. 1984;70:479 484. 20. Larbuisson R, Venneman I, Stiels B. The efficacy and safety of intravenous propafenone versus intravenous amiodarone in the conversion of atrial fibrillation or flutter after cardiac surgery. J Cardiothorac Vasc Anesth. 1996;10:229 234. 21. Mulay A, Kirk AJ, Angelini GD, et al. Posterior pericardiotomy reduces the incidence of supraventricular arrhythmias following coronary artery bypass surgery. Eur J Cardiothorac Surg. 1995;9:150 152. 22. Ascione R, Lloyd CT, Underwood MJ, et al. On pump versus off pump coronary revascularization: evaluation of renal function. Ann Thorac Surg. 1999;68:493 498. 23. Brasil LA, Gomes WJ, Salomao R, et al. Inflammatory response after myocardial revascularization with or without cardiopulmonary bypass. Ann Thorac Surg. 1998;66:56 59. 24. Ascione R, Lloyd CT, Underwood MJ, et al. Inflammatory response after coronary revascularization with or without cardiopulmonary bypass: a prospective randomised study. Ann Thorac Surg. 2000;69:1198 1204. 25. Ascione R, Lloyd CT, Underwood MJ, et al. Economic outcome of off pump coronary artery bypass surgery: a prospective randomised study. Ann Thorac Surg. 1999;68:2237 2242. 26. Tchervenkov CI, Wynands JE, Symas JF, et al. Electrical behavior of the heart following high-potassium cardioplegia. Ann Thorac Surg. 1983;36: 314 319. 27. Smith PK, Buhrman WC, Levett JM, et al. Supraventricular conduction abnormalities following cardiac operations: a complication of inadequate atrial preservation. J Thorac Cardiovasc Surg. 1983;85:105 115. 28. Buckberg GD. Update on current techniques of myocardial protection. Ann Thorac Surg. 1995;60:805 814. 29. Fuller JA, Adams GG, Buxton B. Atrial fibrillation after coronary artery bypass grafting: is it a disorder of the elderly? J Thorac Cardiovasc Surg. 1989;97:821 825. 30. Lemmer JH, Dilling EW, Morton JR, et al. Aprotinin for primary coronary artery bypass grafting: a multicenter trial of three dose regimens. Ann Thorac Surg. 1996;62:1659 1668. 31. Bollmann A, Sonne K, Esperer HD, et al. Non invasive assessment of fibrillary activity in patients with paroxysmal and persistent atrial fibrillation using the Holter ECG. Cardiovasc Res. 1999;44:60 66.