Different results are reported for surgical myocardial

Surgical Revascularization for Acute Coronary Insufficiency: Analysis of Risk Factors for Hospital Mortality Biagio Tomasco, MD, Antonino Cappiello, MD, Rosario Fiorilli, MD, Archimede Leccese, MD, Raniero Lupino, MD, Antonio Romiti, MD, and Ugo F. Tesler, MD Divisions of Cardiac Surgery and Cardiology, Ospedale San Carlo, Potenza, Italy Background. A retrospective study of 444 patients undergoing urgent and emergent coronary artery bypass grafting for acute coronary insufficiency was performed to identify the risk factors for hospital death specifically associated with the clinical severity of the acute coronary insufficiency syndrome. Methods. The patients were divided into three groups urgent, emergent A, and emergent B on the basis of the evolution of the clinical pattern of the acute coronary insufficiency syndrome on full medical treatment. The three categories were defined as follows: urgent (257 patients), surgical revascularization could be delayed for 24 to 36 hours after surgical consultation because of adequate control of ischemia; emergent A (127 patients), prompt myocardial revascularization was required because medical treatment achieved only transient regression of an unrelenting ischemic pattern; and emergent B (60 patients), prompt myocardial revascularization was required because the acute coronary insufficiency was entirely refractory to medical treatment. Results. Mortality rates were 7.4% for the urgent group, 13.4% for the emergent A group, and 31.7% for the emergent B group. Multivariate analysis identified the following as risk factors for hospital mortality: ejection fraction (p 0.023) and aortic cross-clamp time (p 0.10) for the urgent group; aortic cross-clamp time (p 0.017), ejection fraction (p 0.03), and nonuse of blood cardioplegia (p 0.04) for the emergent A group; and cardiogenic shock (p 0.00), preoperative ischemic interval (p 0.00), aortic cross-clamp time (p 0.018), and nonuse of blood cardioplegia (p 0.012) for the emergent B group. Conclusions. A more exact definition of patient risk can be achieved when predictive outcome models are constructed using the risk factors specifically related to each level of clinical severity of the ischemic syndrome. (Ann Thorac Surg 1997;64:678 83) 1997 by The Society of Thoracic Surgeons Different results are reported for surgical myocardial revascularization performed for acute coronary insufficiency (ACI). The differences, related at least in part to the wide spectrum of severity in clinical presentation of the ACI syndrome, are most probably explained by the lack of uniform criteria identifying comparable subsets of patients. This study was performed to determine a more reliable prognostic identification of the surgical risk in patients with different levels of severity of the ACI syndrome. Material and Methods Definitions Angiographic morphology of the culprit coronary lesion was described in accordance with the criteria of Freeman and colleagues [1]. Preoperative ischemic interval was defined as the interval between the onset of the clinical or the electrocardiographic signs or both of ACI and the Accepted for publication March 7, 1997. Address reprint requests to Dr Tomasco, Division of Cardiac Surgery, Ospedale San Carlo, 85100 Potenza, Italy. institution of cardiopulmonary bypass. The degree of severity of the coronary artery disease was evaluated by applying the ischemic score obtained by the wall motion abnormality for each coronary stenosis [2]. Myocardial revascularization was considered complete when all stenoses with a score higher than 1 were bypassed [2]. Recent myocardial infarction was considered one occurring within 3 weeks prior to revascularization. Cardiogenic shock was defined as the clinical state of hypoperfusion characterized by systolic pressure lower than 80 mm Hg and central filling pressure greater than 20 mm Hg or cardiac index of less than 1.8 L min 1 m 2. Low cardiac output was considered present when clinical signs of hypoperfusion were associated with elevated central venous and pulmonary capillary pressures, mean systemic pressure of less than 70 mm Hg, cardiac index lower than 2.0 L min 1 m 2, and metabolic acidosis. The diagnosis of perioperative myocardial infarction was made on the basis of the presence of at least two of the following: new Q waves on the electrocardiogram, values of the MB fraction greater than 10% of the total creatine kinase values, and new left 1997 by The Society of Thoracic Surgeons 0003-4975/97/$17.00 Published by Elsevier Science Inc PII S0003-4975(97)00541-9

Ann Thorac Surg TOMASCO ET AL 1997;64:678 83 ACUTE CORONARY INSUFFICIENCY 679 Table 1. Preoperative Profiles of the Three Patient Groups (n 257) (n 127) Emergent B (n 60) Age (y) b 60.4 7.5 61.5 7.7 59.5 8.8 Sex (%) Male 81.7 66.9 88.3 Diabetes (%) 71.4 69.2 61.5 Recent myocardial 9.3 13.4 infarction (%) Reoperation (%) 1.2 2.4 Ejection fraction a 0.57 0.13 0.58 0.14 0.54 0.14 Ischemic score a 8.7 2.6 8.13 2.8 8.22 2.4 Ischemia-producing lesion Complex 128 76 21 Not complex 68 20 Thrombus 10 9 39 Not identified 51 22 Preoperative ischemic...... 3.91 2.4 interval (h) a Cardiogenic shock (%)...... 45% a Data are shown as the mean standard deviation. Reprinted from Tomasco B, Cappiello A, Fiorilli R, et al. Rivascolarizzazione chirurgica dell insufficienza coronarica acuta: analisi dei fattori di rischio di mortalità ospedaliera in urgenza ed emergenza. G Ital Cardiol 1995;25:269 80; by permission of Piccin Nuova Libraria S.p.A. ventricular akinetic areas in the postoperative echocardiogram. Full medical treatment included intravenous administration of nitroglycerin, calcium antagonist, and heparin sodium. Patient Population Between January 1, 1985, and December 31, 1992, a consecutive series of 444 patients underwent surgical myocardial revascularization for ACI. Patients were divided into three groups on the basis of the preoperative clinical and electrocardiographic evolution patterns of the ischemic syndrome. The first group, the urgent group, comprised 257 patients in whom full medical treatment achieved adequate control of ischemia. Generally, these patients had operation within 24 to 36 hours after surgical consultation because subsequent mild episodes of recurrent ischemia were thought not to require emergent revascularization. The second group, the emergent A group, consisted of 127 patients who required prompt myocardial revascularization because of inadequate control of ischemia. An unrelenting ischemic pattern, interrupted by transient periods of regression, was the usual feature characterizing this group of patients. The third group, the emergent B group, comprised 60 patients with ongoing ischemia. It lasted from the inception of the clinical presentation of the syndrome or, less frequently, persisted for at least 30 minutes before the institution of cardiopulmonary bypass. Demographics and clinical characteristics of each group are summarized in Table 1. In particular, the coronary angiogram revealed that thrombus was the ischemia-producing lesion in 58 patients (10 in the urgent group, 9 in the emergent A group, and 39 in the emergent B group); 19 of these patients received thrombolytic treatment. In 73 patients with diffuse coronary artery disease (22 in the emergent A group and 51 in the urgent group), the culprit lesion could not be identified, even when electrocardiographic or wall motion abnormalities were considered. Regarding noncardiac comorbidity, the incidences of diabetes (22%), chronic obstructive pulmonary disease (9%), renal failure (2.6%), and previous cerebral vascular accident (4%) were closely matched in the three groups. Patients with ACI secondary to the failure of percutaneous transluminal coronary angioplasty were excluded from this study for two reasons: (1) because of the favorable lesions seen in these patients in terms of reduced severity and extension of coronary artery disease and (2) because the patients had been included in a protocol allowing the possibility of emergent surgical revascularization. Surgical Techniques Operative data are given in Table 2. Cardiopulmonary bypass was instituted using an ascending aortic cannula and a two-stage single venous cannula. Moderate hemodilution (hematocrit, 20% to 25%), moderate systemic hypothermia (28 C), and a flow of 2.5 L min 1 m 2 were maintained. Three types of myocardial protection were employed during the study period: St. Thomas I crystalloid cardioplegic solution [3] was used in 298 patients in urgent and emergent A groups and in 36 patients in the emergent B group; blood cardioplegia with cold induction [4] was used in 86 patients in the urgent and emergent A groups; and blood cardioplegia with warm induction and substrate enrichment [4] was used in 24 patients in the emergent B group. The cardioplegic solutions were administered in an antegrade fashion through the aortic root and through the proximal ends of the grafts. Table 2. Operative Data for the Three Patient Groups (n 257) (n 127) Emergent B (n 60) Distal anastomosis b 2.87 1.1 2.61 1.1 2.47 0.8 Completeness of 68 61.3 65.5 revascularization (%) Endarterectomy 16 15.7 13.3 (%) Aortic clamp time 51.1 21.4 43.1 21.1 43 20.5 (min) b Internal mammary 18.8 6.3 5 artery graft (%) Blood cardioplegia (%) 22.2 22.8 40 Reprinted from Tomasco B, Cappiello A, Fiorilli R, et al. Rivascolarizzazione chirurgica dell insufficienza coronarica acuta: analisi dei fattori di rischio di mortalità ospedaliera in urgenza ed emergenza. G Ital Cardiol 1995;25:269 80; by permission of Piccin Nuova Libraria S.p.A.

680 TOMASCO ET AL Ann Thorac Surg ACUTE CORONARY INSUFFICIENCY 1997;64:678 83 Table 3. Operative Mortality and Morbidity a,b (n 257) (n 127) Emergent B (n 60) p Value ( 2 ) Mortality 7.4 13.4 31.7 0.001 Perioperative 11 11.8 63.3 0.001 infarction Low output 8.3 9.7 33.3 0.001 syndrome Respiratory 20.2 16.1 31.3 0.08 failure Neurologic 2.8 3.4 10.4 0.04 deficit Intraaortic balloon pump 1.9 4 16.9 0.001 0.2 b a Data are shown as percentages. b Significance: emergent A versus urgent. Reprinted from Tomasco B, Cappiello A, Fiorilli R, et al. Rivascolarizzazione chirurgica dell insufficienza coronarica acuta: analisi dei fattori di rischio di mortalità ospedaliera in urgenza ed emergenza. G Ital Cardiol 1995;25:269 80; by permission of Piccin Nuova Libraria S.p.A. Statistical Analysis The 2 test was used to detect any significant correlation between outcome and surgical priority. Multivariate logistic regression analysis was used to select the model of variables with the highest predictive power for the event death. Only variables with a p value of less than 0.05 for improvement in 2 were included in the model. All calculations were made using a standard statistical program (BMDP Statistical Software, Los Angeles, CA). Results Mortality and Morbidity Hospital mortality (55 patients) and morbidity are shown in Table 3. Death resulted from cardiac-related causes in 43 patients: 11 died of myocardial failure, 31 of low cardiac output syndrome (26 with related myocardial infarction), and 1 patient of a major sudden irreversible rhythm disorder. In addition, 3 patients requiring prolonged assisted ventilation died of pulmonary causes, 5 died of neurologic causes, 3 died of gastrointestinal bleeding, and 1 patient died of sepsis. Predictors of Operative Death Table 4 shows the variables in the statistical analysis for the urgent and emergent A groups. The following risk variables were significant: ejection fraction, aortic crossclamp time, surgical priority, and recent myocardial infarction. A change in surgical priority from urgent to emergent A increased the probability of death especially in patients with a lower ejection fraction or longer aortic cross-clamp time (Figs 1, 2). To examine the hypothesis that the variable nonuse of blood cardioplegia could have a different effect in the two subgroups (blood cardioplegia and crystalloid cardioplegia) defined according to patient surgical priority, we carried out a separate analysis for each group. This revealed that the variable had a negative impact on hospital mortality only in the emergent A group (Table 5). The risk factors identified as predictors of death in the emergent B group are listed in Table 6. The combination of cardiogenic shock and prolonged preoperative ischemic interval ( 6 hours) had the highest predictive value despite use of blood cardioplegia (Fig 3). Comment Mortality rates for surgical myocardial revascularization in patients with ACI range from 2.5% [5] to 15.6% [6]. Phillips and associates [7] reported an operative mortality rate of 5.7%. Included in their study were patients who underwent myocardial revascularization within 36 hours of the onset of the ischemic syndrome. At the Cleveland Clinic [6], Golding and colleagues [5] obtained a mortality rate of 2.5% for emergency operations after failure of Table 4. Logistic Analysis of Risk Factors for Hospital Death in and Groups Univariate Analysis Multivariate Regression Analysis Stepwise Selection p Value Coefficient ( 1 SE) Step p Value ( 2 Improvement) Ejection fraction 0.001 0.31 10 1 (0.13 10 1 ) 1 0.002 Aortic cross-clamp time 0.008 0.21 10 1 (0.86 10 2 ) 2 0.020 revascularization 0.06 0.86 (0.37) 3 0.017 Recent myocardial infarction 0.003 0.96 (0.46) 4 0.049 Ischemic score 0.04 0.34 10 1 (0.79 10 1 ) Age NS Sex NS Endarterectomy NS...... Nonuse of blood cardioplegia NS... Incomplete revascularization NS...... Constant... 2.44 (1.18) NS not significant; SE standard error. Reprinted from Tomasco B, Cappiello A, Fiorilli R, et al. Rivascolarizzazione chirurgica dell insufficienza coronarica acuta: analisi dei fattori di rischio di mortalità ospedaliera in urgenza ed emergenza. G Ital Cardiol 1995;25:269 80; by permission of Piccin Nuova Libraria S.p.A.

Ann Thorac Surg TOMASCO ET AL 1997;64:678 83 ACUTE CORONARY INSUFFICIENCY 681 Fig 1. Probability of hospital death in emergent A and urgent groups according to preoperative ejection fraction. Dotted lines indicate 70% confidence intervals. (Reprinted from Tomasco B, Cappiello A, Fiorilli R, et al. Rivascolarizzazione chirurgica dell insufficienza coronarica acuta: analisi dei fattori di rischio di mortalità ospedaliera in urgenza ed emergenza. G Ital Cardiol 1995;25:269 80; by permission of Piccin Nuova Libraria S.p.A.) Fig 2. Probability of hospital death in emergent A and urgent groups according to aortic cross-clamp time. Dotted lines indicate 70% confidence intervals. (Reprinted from Tomasco B, Cappiello A, Fiorilli R, et al. Rivascolarizzazione chirurgica dell insufficienza coronarica acuta: analisi dei fattori di rischio di mortalità ospedaliera in urgenza ed emergenza. G Ital Cardiol 1995;25:269 80; by permission of Piccin Nuova Libraria S.p.A.) percutaneous transluminal coronary angioplasty, but a considerable number of surgical revascularization procedures were performed within 24 hours after the onset of ischemia. Teoh and coauthors [8] reported an operative mortality rate of 8.5% for urgent surgical revascularization. Classified as urgent were all patients with unstable angina operated on within 72 hours of catheterization. There was no mention of the impact of emergency operation on mortality. Fremes and associates [9] found an operative mortality rate of 9.2% for patients with unstable angina who needed urgent revascularization. Those authors stated, however, that certain important characteristics have not been investigated throughout the study period such as emergent as opposed to urgent revascularization. The markedly higher mortality rate of 14.5% reported by Edwards and colleagues [10] was explained by them as a consequence of the exclusion of all patients in whom revascularization had been deferred up to 24 hours after surgical consultation: only patients with ongoing ischemia were considered to need emergent coronary artery bypass grafting, but no mention was made about preoperative hemodynamic conditions. Therefore, to categorize comparable subsets of patients and to detect a more reliable prognostic stratification of the surgical risk, uniform criteria identifying different levels of severity of ACI syndrome need to be defined. In our study, when ACI was adequately controlled, although not completely, patients were considered as representing the first degree of clinical severity and consequently were grouped as urgent. Patients showing an ischemic pattern characterized by unrelenting ischemia interrupted by periods of transient regression were con- Table 5. Logistic Analysis of Risk Factors for Hospital Death in Versus Revascularization p Value ( 2 improvement) Coefficient ( 1 SE) p Value ( 2 improvement) Coefficient ( 1 SE) Ejection fraction 0.023 0.36 10 1 (0.16 10 1 ) Aortic cross-clamp 0.017 0.29 10 1 (0.12 10 1 ) time Aortic crossclamp 0.10 0.17 10 1 (0.10 10 1 ) Ejection fraction 0.033 0.38 10 1 (0.19 10 1 ) time Nonuse of blood 0.04 1.77 (1.07) cardioplegia Constant... 1.47 (1.08) Constant... 0.85 (1.21) SE standard error. Reprinted from Tomasco B, Cappiello A, Fiorilli R, et al. Rivascolarizzazione chirurgica dell insufficienza coronarica acuta: analisi dei fattori di rischio di mortalità ospedaliera in urgenza ed emergenza. G Ital Cardiol 1995;25:269 80; by permission of Piccin Nuova Libraria S.p.A.

682 TOMASCO ET AL Ann Thorac Surg ACUTE CORONARY INSUFFICIENCY 1997;64:678 83 Table 6. Logistic Analysis of Risk Factors for Hospital Death in Emergent B Group Univariate analysis Multivariate Regression Analysis p Value Coefficient ( 1 SE) Step Stepwise Selection p Value ( 2 Improvement) Cardiogenic shock 0.000 4.9 (1.7) 1 0.0001 Preoperative ischemic interval 6 hours 0.002 3.8 (1.4) 2 0.005 Nonuse of blood cardioplegia NS 3.1 (1.5) 3 0.012 Aortic cross-clamp time 0.000 0.60 10 1 (0.47 10 1 ) 4 0.018 Ischemic score 0.04 0.38 10 1 (0.37 10 1 ) Female sex 0.015 2.04 (2.44)... Incomplete revascularization 0.07 0.48 (1.5) Age NS Ejection fraction NS... Constant... 6.5 (3.7)... NS not significant; SE standard error. Reprinted from Tomasco B, Cappiello A, Fiorilli R, et al. Rivascolarizzazione chirurgica dell insufficienza coronarica acuta: analisi dei fattori di rischio di mortalità ospedaliera in urgenza ed emergenza. G Ital Cardiol 1995;25:269 80; by permission of Piccin Nuova Libraria S.p.A. sidered as belonging to the second degree of clinical severity and were grouped as emergent A. Finally, patients with ongoing ischemia lasting throughout the preoperative period or for at least 30 minutes before the institution of cardiopulmonary bypass were considered as belonging to the third degree of clinical severity and grouped as emergent B. The validity of this stratification is confirmed by the different mortality rates observed (see Table 3). Moreover, in regard to the first two groups, the risk factors prolonged aortic cross-clamp time and low ejection fraction carry a significantly worse prognostic weight for the emergent A group than the urgent group (see Figs 1, 2). Also, the risk factor nonuse of blood cardioplegia increases mortality exclusively in the emergent A group (see Table 5). It can be inferred that within this last group, recent repeated ischemic episodes, albeit brief, may cause prolonged metabolic alterations [11], thus increasing myocardial vulnerability; therefore, the better results obtained with the use of blood cardioplegia may be related to improvement in metabolic conditions (Fig 4). Statistical analysis of the emergent B group allowed the identification of three subgroups of patients with different surgical risk: patients not in cardiogenic shock, patients in cardiogenic shock and with a preoperative ischemic interval shorter than 6 hours, and patients in cardiogenic shock and with a preoperative ischemic interval longer than 6 hours. Although there are contrasting conclusions regarding the importance of the preoperative ischemic interval [6, 7, 12 15], our experience demonstrates that in patients in cardiogenic shock, early surgical revascularization performed with the use of blood cardioplegia allows us to expect a mortality rate of 10% (see Fig 3). In patients in cardiogenic shock who are operated on after an ischemic interval longer than 6 hours, our expected mortality rate appears to be greater than 50% despite the use of blood cardioplegia (see Fig 3). Better results would probably have been obtained by applying advanced techniques of myocardial protection [16 18], as they permit a more complete recovery of the myocardial mass. Adopting prolonged regional controlled reperfusion, Allen and co-workers [19] reported a reduction in mortality rate from 17% to 3%. In conclusion, granted that the exact definition of patient risk factors and the identification of new risk factors make the interpretation of the outcome data Fig 3. Probability of hospital death in emergent B group according to cardiogenic shock, cardioplegia, and preoperative ischemic interval. Each triangle represents the probability of death on the basis of the possible combinations of risk factors shown below the abscissa. Each of the combinations is converted to a score (shown on abscissa). (B blood; C crystalloid.) (Reprinted from Tomasco B, Cappiello A, Fiorilli R, et al. Rivascolarizzazione chirurgica dell insufficienza coronarica acuta: analisi dei fattori di rischio di mortalità ospedaliera in urgenza ed emergenza. G Ital Cardiol 1995;25:269 80; by permission of Piccin Nuova Libraria S.p.A.)

Ann Thorac Surg TOMASCO ET AL 1997;64:678 83 ACUTE CORONARY INSUFFICIENCY 683 Fig 4. Probability of hospital death in emergent A subgroups: blood cardioplegia and crystalloid cardioplegia according to preoperative ejection fraction. Dotted lines indicates 70% confidence intervals. (Reprinted from Tomasco B, Cappiello A, Fiorilli R, et al. Rivascolarizzazione chirurgica dell insufficienza coronarica acuta: analisi dei fattori di rischio di mortalità ospedaliera in urgenza ed emergenza. G Ital Cardiol 1995;25:269 80; by permission of Piccin Nuova Libraria S.p.A.) clearer, our study supports three major conclusions. First, the exact definition of urgent surgical priority excludes patients undergoing semielective revascularization procedures. Including them would lead to better results, thus giving the impression of low mortality rates in a high-risk category of patients. Second, in ACI syndrome, emergent A surgical priority must be considered a new and important risk factor in the assessment of the exact risk-adjusted patient outcome. Including patients with emergent A surgical priority with high-risk patients undergoing emergency bypass would permit a lower riskadjusted mortality. Third, results related to the various degrees of severity of the ACI syndrome must be weighed against the type of myocardial protection employed. This study was supported in part by a grant from the Italian National Council for Research. We thank Dr Gerald M. Lemole for his encouragement and his helpful advice in the preparation of this report. References 1. Freeman MR, Williams AE, Chisholm RJ, Armstrong PW. Intracoronary thrombus and complex morphology in unstable angina. Relation to timing of angiography and inhospital cardiac events. Circulation 1989;80:17 23. 2. Kirklin JW, Barratt-Boyes BG, eds. Cardiac surgery. 1st ed. New York: Churchill Livingstone, 1986:253, 263. 3. Hearse DJ, Brainbridge MV, Jynge P, eds. Cardioplegia. 1st ed. New York: Raven, 1981:346 8. 4. Rosenkranz EL, Buckberg GD, Laks H, Mulder DG. Warm induction of cardioplegia with glutamate-enriched blood in coronary patients with cardiogenic shock who are dependent on inotropic drugs and intra-aortic balloon support. J Thorac Cardiovasc Surg 1983;86:507 18. 5. Golding LAR, Loop FD, Hollman JL, et al. Early results of emergency surgery after coronary angioplasty. Circulation 1986;74(Suppl 3):26 9. 6. Athanasuleas CL, Geer DA, Arciniegas JG, et al. A reappraisal of surgical intervention for acute myocardial infarction. J Thorac Cardiovasc Surg 1987;93:405 14. 7. Phillips SJ, Zeff RH, Skinner JR, Toon RS, Grignon A, Kongtahworn C. Reperfusion protocol and results in 738 patients with evolving myocardial infarction. Ann Thorac Surg 1986;41:119 25. 8. Teoh KH, Christakis GT, Weisel RD, et al. Increased risk of urgent revascularization. J Thorac Cardiovasc Surg 1991;93: 291 9. 9. Fremes SE, Goldman BS, Christakis GT, et al. Current risk of coronary bypass for unstable angina. Eur J Cardio-thorac Surg 1991;5:235 43. 10. Edwards FH, Bellamy RF, Burge JR, et al. True emergency coronary artery bypass surgery. Ann Thorac Surg 1990;49: 603 11. 11. Braunwald E, Kroner RA. The stunned myocardium: prolonged post-ischemic ventricular dysfunction. Circulation 1982;66:1146 9. 12. Berg R, Selinger SL, Leonard JI, et al. Acute myocardial infarction a surgical emergency. J Thorac Cardiovasc Surg 1984;88:902 6. 13. DeWood MA, Spores J, Berg R, et al. Acute myocardial infarction: a decade of experience with surgical reperfusion in 701 patients. Circulation 1983;68(Suppl 2):8 16. 14. Flameng W, Sergeant P, Van Haeche J, Suy R. Emergency coronary bypass grafting for evolving myocardial infarction. J Thorac Cardiovasc Surg 1987;94:124 31. 15. Buckberg GD. Strategies and logic of cardioplegic delivery to prevent, avoid and reverse ischemic and reperfusion damage. J Thorac Cardiovasc Surg 1987;93:127 39. 16. Allen BS, Buckberg GD, Schwaiger M, et al. Studies of controlled reperfusion after ischemia. XVI. Early recovery of regional wall motion in patients following surgical revascularization after eight hours of acute coronary occlusion. J Thorac Cardiovasc Surg 1986;92:636 48. 17. Beyersdorf F, Mitrev Z, Sarai K, et al. Changing patterns of patients undergoing emergency surgical revascularization for acute coronary occlusion. Importance of myocardial protection technique. J Thorac Cardiovasc Surg 1993;106: 137 48. 18. Beyersdorf F, Acar C, Buckberg GD, et al. Studies on prolonged regional ischemia. III. Early natural history and simulated single and multivessel disease with emphasis on remote myocardium. J Thorac Cardiovasc Surg 1989;98: 368 80. 19. Allen BS, Buckberg GD, Fontan FM, et al. Superiority of controlled surgical reperfusion versus percutaneous transluminal coronary angioplasty in acute coronary occlusion. J Thorac Cardiovasc Surg 1993;105:864 84.