Ann Thorac Surg 2000;70:807-812
© 2000 The Society of Thoracic Surgeons
Original articles: cardiovascular
Reperfusion injury associated with one-fourth of deaths after coronary artery bypass grafting
Saara M. Weman, MDb,
Pekka J. Karhunen, MDc,d,
Antti Penttilä, MDa,
Antero A. Järvinen, MD, PhDb,
Ulla-Stina Salminen, MDb
a Department of Forensic Medicine, Helsinki University, Helsinki, Finland
b Department of Surgery, Helsinki University Hospital, Helsinki, Finland
c School of Medicine, University of Tampere, Tampere, Finland
d Department of Clinical Pathology and Forensic Medicine, University of Kupio, Kupio, Finland
Address reprint requests to Dr Salminen, Department of Surgery, PL 340, Helsinki University Hospital, Haartmaninkatu 4, Helsinki, FIN-00029 HYKS, Finland
e-mail: ulla.s.salminen{at}netsonic.fi
 |
Abstract
|
|---|
Background. This study of reperfusion injury after coronary artery bypass grafting focuses on its contribution to fatal outcome, on its connection with myocardial infarction (MI) and on risk factors.
Methods. A consecutive series of 190 patients (mean age 61.7 ± 8.9 years) dying within 30 days following coronary artery bypass grafting was autopsied with concomitant postmortem angiography during 1980 to 1993.
Results. Reperfusion injury was revealed in 49 (25.8%) patients, with concomitant MI in almost all (46 of 49) (p < 0.01). Reperfusion injury occurred in association with preoperative New York Heart Association (NYHA) III classification (p < 0.05), coronary endarterectomy (p < 0.01), long aortic clamping time (p < 0.01), and short postoperative survival (p < 0.05).
Conclusions. Reperfusion injury was observed in one fourth of the deaths in association with MI. It occurred more often in patients with preoperative NYHA III symptoms and in those in whom endarterectomy was carried out and the anoxic time of the myocardium was longer. The shorter postoperative survival time indicates the lethal nature of this complication.
|
Introduction
|
|---|
Reperfusion injury (RI) syndrome is a well-known phenomenon following restoration of the coronary circulation after coronary artery bypass grafting (CABG) [1, 2]. Clinically, it is characterized by arrhythmias combined with myocardial and microvascular stunning [3], often leading to low cardiac output syndrome [3, 4]. Myocardial infarction (MI) may be indistinguishable from RI, and their association has been noted in patients with fatal outcome [1].
The morphologic counterpart of RI is a special form of myocardial necrosis known as contraction band necrosis [5], which has been present in 31% to 100% of CABG patients with fatal outcome [1, 6, 7]. No reports focusing on clinical risk factors predicting RI have been found in the literature.
The aim of this study was to determine frequency of and risk factors for RI and its connection with preoperative or intraoperative MI in a prospective autopsy series of 190 patients who died within 30 days following CABG. We also assessed the value of a novel method of postmortem angiography in the diagnosis of myocardial injury.
|
Patients and methods
|
|---|
The study was approved by the institutional ethics committees. The original prospective series comprised 252 consecutive patients who had a fatal outcome within 30 days after CABG from 1980 to 1993 in all three cardiac centers (Helsinki University Central Hospital, the Deaconess Hospital and the Mehiläinen Hospital) in Helsinki. During this time the total of CABG procedures was 8,286. Autopsies were performed on all those with fatal outcome at the Department of Forensic Medicine, University of Helsinki. Postmortem angiographic roentgenograms taken in connection with the autopsies, as well as the microscopic specimens of myocardium were available in 190 (76.0%) cases. The mean age of the 50 (26%) female patients was 64.9 ± 8.3 years; for the 140 (74%) males, the mean age was 60.5 ± 8.8 years. Mean survival time was 83.4 ± 147.0 hours (range 0 to 678.5 hours), and mean time from death to autopsy was 6.0 ± 1.8 days.
Clinical data relevant for evaluation of the risk factors (Table 1), as well as the operative (Table 2) and postoperative (Table 3) course of each patient, were analyzed retrospectively from hospital records. Of the patients, 33 (17.4%) underwent emergency operation, due to preoperative MI (18 patients), unstable angina [8], percutaneous transluminal coronary angioplastic complication [3], coronary angiographic complication [3], or ventral septal defect following MI [1]. Of the 23 (12.1%) patients with preoperative MI, 13 (56.5%) had had cardiogenic shock; 7 (53.8%) of these 13 received preoperative treatment with the intraaortic balloon pump.
Cardiopulmonary bypass was used routinely. Cold crystalloid high potassium cardioplegia (St. Thomas) was used for 186 (97.9%) patients for myocardial protection during aortic clamping. Four (2.1%) operations were performed with a beating heart. Concomitant cardiac procedures, the most frequent of which were coronary endarterectomies, were carried out in 78 (41.1%) patients, and vascular procedures in 4 (2.1%) (Table 4). The mean number of bypasses was 3.8 ± 1.1 (range 1 to 7), with 43 (22.6%) patients dying intraoperatively. Mean postoperative survival time for the remaining 147 (77.4%) was 118.1 ± 150.5 hours (range 1.3 to 678.4 hours).
View this table:
[in this window]
[in a new window]
|
Table 4. Concomitant Cardiac and Vascular Procedures for the 190 Study Patients With Fatal Outcome Following CABG
|
|
Postmortem angiographic method
At autopsy, the chest was opened and the sternum removed. The brachiocephalic trunk or the abdominal aorta was cannulated, and the axillary arteries were clamped distal to the base of the internal thoracic arteries. Liquid silicone rubber (1,000 g) was made radiopaque with lead oxide (210 g) and mixed with solidifier (20 mL), and immediately the grafts and coronary arteries were perfused with this mixture through the cannula at a physiologic pressure of 120 to 180 mm Hg. This mixture solidifies to a cast in 2 to 3 hours. Thereafter, roentgenograms were taken. Areas of filling defects suggestive of myocardial injury were recorded retrospectively, independent of any knowledge of autopsy results. This method was described in detail elsewhere [8].
Histologic criteria
In reevaluation of microscopic examination, reperfusion injury was characterized by the appearance of contraction band necrosis of myocytes of the left ventricle. The area of this necrosis was scored in each case semiquantitatively (0 to 3+), using both phosphotungstic acid hematoxylin (PTAH) and Herowic staining. The score was 0 when no contraction band was observed in histologic sections of the myocardium, 1+ for contraction bands in scattered myocytes, 2+ when scattered myocytes showed contraction bands along with small areas of contraction band necrosis, and 3+ when contraction bands were detected in larger areas of the myocardium. Fresh MI was defined conventionally by the appearance of polymorphonuclear leukocytes.
Statistics
Data are expressed as mean ± SD. To study the association of RI and its possible risk factors, we used logistic regression analysis, in which the presence or absence of RI was the dependent factor with various risk factors as the independent factors. The categorical variables were coded as dummy (yes/no) variables. The computation was performed with the SPSS/Win statistical software (Version 9.0). The Pearson’s 
-square test with Yates correction was used in comparing the annual decrease in the frequency of RI.
|
Results
|
|---|
The mortality during 1980 to 1993 was 3.0%. Mortality was highest in 1983 (5.2%) and decreased thereafter year by year, being 1.7% during both the last years of the study.
Frequency of reperfusion injury
Of our 190 patients, 49 (25.8%) showed RI characterized by contraction bands in the myocardium. Of these, 12 (24.5%) scored as mild (1+), 9 (18.4%) as moderate (2+), and 28 (57.1%) as severe (3+). Incidence of RI decreased from 44% to 12% between 1980 and 1993 (p < 0.01) (Fig 1). In the group with RI, the mean time between death and autopsy was 5.9 ± 1.7 (range 3 to 10) days and in those with no findings of RI, mean time was 6.0 ± 1.9 (range 2 to 12) days.
View larger version (40K):
[in this window]
[in a new window]
|
Fig 1. The frequency of reperfusion injury (RI) significantly (p < 0.01) decreased during a period of 14 years in 190 patients dying within 30 days after coronary artery bypass grafting. Mortality (%) and the number of autopsies (N) during each time period are given in numbers.
|
|
Association of reperfusion injury with preoperative characteristics
Patients with or without RI did not differ significantly in preoperative characteristics (Table 1), except that RI was more frequent in those with New York Heart Association (NYHA) III symptoms (p < 0.05). The number of diseased coronary arteries did not affect the frequency of RI. In those with RI, one-vessel disease was found in 1 (2.0%) patient, two-vessel disease in 5 (10.2%), and three-vessel disease in 42 (85.7%) patients. In patients with no RI, one-, two-, and three-vessel disease was found in 4 (2.9%), 14 (10.2%), and 119 (86.9%) patients, respectively. A stenosis of 95% or greater was found and bypassed in 39 (80.0%) patients with RI. The mean number of these stenoses was 1.53 ± 1.7 (range 0 to 5). Of patients with no RI, 112 (79.4%) had stenoses of 95% or greater, the number of which was 1.53 ± 1.7 (range 0 to 5). Within the previous 30 days, 23 (12.1%) had experienced preoperative MI. Reperfusion injury was detected postoperatively in only 4 (17.4%) of these, all 4 of whom had developed MI within a period of 3 days or less before the operation. Preoperative cardiogenic shock was seen in 3 (6.1%) patients with postoperative RI and in 10 (7.1%) patients with none.
Operative risk factors and postoperative events associated with reperfusion injury
In the 49 patients with RI, aortic clamping time (p < 0.01) was longer than in those with no contraction band necrosis in microscopic examination (Table 2). A correlation existed between severity of RI score and aortic clamping time (Table 5). Coronary endarterectomy appeared to be a risk factor (Table 4), being twice as frequent (34.7%) in patients with a finding of contraction band necrosis as in patients without (p < 0.01). The period of postoperative shock was 2,787 ± 3,720 minutes (range 95 to 16,850 minutes) in patients with RI, and 3,977 ± 7,304 minutes (range 22 to 36,451 minutes) in those with no RI. Mean postoperative survival time of patients with RI (59.1 hour) was significantly shorter (p < 0.05) than in patients without RI (135.8 hours) (Tables 3 and 5). In patients with RI, the cause of death was almost always cardiac (Table 6).
View this table:
[in this window]
[in a new window]
|
Table 5. Correlation of Operative and Postoperative Events (Mean ± SD) With Reperfusion Injury, Separately for a Group With Findings of Grade 1 and Grade 2 Reperfusion Injury and a Group With Grade 3
|
|
Correlation of reperfusion injury with operative myocardial infarction
Reperfusion injury showed a correlation with histologically confirmed acute MI (Tables 3 and 5). Of the 49 patients with RI, 46 (93.9%) also had acute MI, with MI less frequent (p < 0.01) in those with no RI (72.3%). Frequency of old MI scars did not differ between patients with (85.7%) or without (79.4%) RI.
Resuscitation
Of patients with RI, 1 (2.0%) was resuscitated preoperatively and 17 (34.7%) were resuscitated postoperatively. Of those with no RI, preoperative resuscitation was given in 6 (4.3%) and postoperative resuscitation in 58 (41.1%) cases.
Postmortem angiography
In postmortem angiography, 102 (53.7%) patients had an angiographic filling defect of the myocardial vasculature (Table 3). A significant correlation (p < 0.01) was reached between myocardial filling defect and RI. Angiographs showed filling defects in 69.4% of patients with RI and in 48.2% without RI (Tables 3 and 5).
|
Comment
|
|---|
Reperfusion injury is a common phenomenon associated with CABG [1, 2]. Events related to this disease comprise arrhythmias, vascular damage, the no-reflow phenomenon, and myocardial functional stunning. Cell swelling and hastening of necrosis in cells already irreversibly damaged by ischemia are also ascribed to RI [3, 4]. The mechanisms and pathogenesis by which the lesion is caused are complex and still uncertain. The two most important hypotheses explaining cellular events involved in RI are calcium overload and free-radical damage [2, 9–11].
In our series of patients with fatal outcome after CABG, our finding of frequency of 25.8% for RI was similar to that of a previous autopsy study [1], which detected RI in 31.0% of 58 patients. Evidence exists that the vascular endothelium becomes severely injured secondary to ischemia and subsequent reperfusion [12–14]. This injury seems to be related to postischemic contractile dysfunction [13], a clinical sign of RI [5]. In the present study, endarterectomy turned out to be one risk factor for RI. In cases of endarterectomy, some mechanical injury, in addition to ischemia, alters the endothelium during the aortic clamping time and results in more severe endothelial damage than in cases with ischemic injury only. However, our finding does not necessarily support the role of endothelial injury in the pathogenesis of RI; it may merely reflect the greater extent of coronary artery disease in these patients with probable preoperative severe endothelial dysfunction. In contrast, although we did not find differences in the number of affected vessels or subtotal or total stenoses, NYHA III symptoms were a preoperative risk factor for RI. These patients have severe clinical ischemic symptoms. In NYHA IV patients no correlation with RI appeared. This finding may be explained by the shorter duration of symptoms and shorter waiting time before operative treatment. However, it was not possible to analyze this finding.
Reperfusion injury was observed in one fourth of the total deaths in our series; it was also associated with 32.7% of our intraoperative deaths, similar to a previous study in which 39.1% of patients with confirmed RI died intraoperatively [1]. In addition, RI limited postoperative survival time. This observation was also made in the study by Bulkley and colleagues [1], in which all but 1 of 15 patients with contraction band necrosis died within 48 hours. In a second autopsy study with RI in all patients, postoperative survival was 1 week or shorter [6]. The connection of RI with both intraoperative death and short postoperative survival time indicates the lethal character of this complication.
Introduction of cold chemical cardioplegic techniques is reported to result in a decrease from 51% to 24% in the rate of regional contraction band necrosis in patients with fatal outcome [7]. Our significant decrease in incidence of RI from 44% to 12% during our study period of 14 years is similar, even though the cardioplegic technique remained unchanged. However, improvement in technical realization as well as in operative and in cardiopulmonary bypass techniques occurred during those years: all factors with an influence on results. This improvement was also observed in the mortality rate. Our finding may also be due at least partly to the fact that the number of annual operations increased and was more than 10 times as high in the end of the study as in the beginning. After introduction of cold chemical cardioplegia, use of intermittent hypothermic cardioplegia with oxygenated blood began in the mid 1980s in some centers [15] or even earlier [16]. The latest development in myocardial protection is continuous warm cardioplegia [15, 17]. This method prevents the deleterious effects of hypothermic ischemia and minimizes RI. The frequency of RI could be expected to be significantly lower in an autopsy series of CABG patients receiving warm blood cardioplegia. However, although warm blood cardioplegia results in less myocardial cell damage, the difference compared with that associated with cold crystalloid cardioplegia is not necessarily significant [18].
In our study, the development of RI was associated with long operative myocardial anoxia. This finding could have been caused by failed myocardial protection during the long period of anoxia. Reperfusion injury was also associated with acute MI, but MI was not a preoperative risk factor, because only 15.5% of MI developed before the operation. However, that all 4 cases with preoperative MI and concomitant RI had developed MI within 3 days before surgery may indicate that patients with MI of only some few days’ duration should be operated on with special care and, based on current knowledge, with continuous warm cardioplegia [15, 17]. Nor did any in our series of patients undergoing CABG with a beating heart develop RI. Evidence exists that selected risk patients should be operated on with a beating heart [19]. With this technique, myocardial cell damage due to the toxic effects of cardioplegic solutions is avoided. Blood circulation is inhibited in each bypassed area only for the time during which the anastomosis is performed. Lymphatic drainage is not affected. However, in our study, the limited number of patients with RI and concomitant preoperative MI, as well as those undergoing CABG with a beating heart, does not allow us to draw conclusions.
In our series, 125 (65.8%) patients had operative MI, 46 (36.8%) of them among the 49 cases with RI. Thus, in almost all of our patients with contraction band necrosis, transmural MI also occurred. Subendocardial operative myocardial injury has been reported to be as high as 82.8%, when transmural injury occurred in only 38% of patients with fatal outcome and a frequency of 31.0% was found in RI [1]. Our correlation of RI with operative MI is in line with this result, although no statistical confirmation of this association has appeared previously [1].
Previously we evaluated the sensitivity and specificity of our postmortem angiographic technique in the diagnosis of graft complications [8]. The present study showed that of the patients with RI, two thirds had a filling defect of the myocardial vasculature on angiographic roentgenograms compared with half of patients without RI. In contrast to MI, which is usually found in myocardial regions of occluded grafts, contraction band necrosis occurs in regions of patent grafts and anastomoses [1, 7]. Our finding is not explained by occluded grafts, but rather by the no-reflow phenomenon related to RI [3]. Cell swelling inhibits blood circulation into the myocardial region and, in postmortem angiography, inhibits spreading of the contrast medium. Thus, postmortem angiography may serve as a preliminary assessment of myocardial injury associated with fatal outcome following CABG.
In conclusion, RI was observed in one fourth of the deaths in our series. Preoperative NYHA III symptoms turned out to be a risk factor. Coronary artery endarterectomy and a long anoxic time were intraoperative risk factors. In association with operative MI, RI resulted in shorter postoperative survival time, indicating the lethal nature of this complication.
|
Acknowledgments
|
|---|
This work has been supported by a research grant from the Finnish Society for Angiology. We are grateful to the staff of the Department of Forensic Medicine. We also thank Professor Ari Harjula from the Deaconess Hospital in Helsinki for his kind help in making hospital records available. We acknowledge Carol Norris, PhD, for the language revision.
|
References
|
|---|
-
Bulkley B.H., Hutchins G.M. Myocardial consequences of coronary artery bypass graft surgery, the paradox of necrosis in areas of revascularization. Circulation 1977;56:906-913.[Abstract/Free Full Text]
-
Ferreira R., Burgos M., Milei J., et al. Effect of supplementing cardioplegic solution with deferoxamine on reperfused human myocardium. J Thorac Cardiovasc Surg 1990;100:708-714.[Abstract]
-
Vaage J., Valen G. Pathophysiology and mediators of ischemia-reperfusion injury with special reference to cardiac surgery. Scand J Thor Cardiovasc Surg Suppl 1993;41:1-18.
-
Opie L.H. Reperfusion injury and its pharmacologic modification. Circulation 1989;80:1049-1062.[Abstract/Free Full Text]
-
Naka Y., Stern D.M., Pinsky D.J. The pathophysiology and biochemistry of myocardial ischemia, necrosis, and reperfusion. In: Fuster V., Ross R., Topol E.J., eds. Atherosclerosis and coronary artery disease. Philadelphia: Lippincott-Raven, 1996:807-814.
-
Hutchins G.M., Bulkley B.H. Correlation of myocardial contraction band necrosis and vascular patency, a study of coronary artery bypass graft anastomosis at branch points. Lab Invest 1977;36:642-648.[Medline]
-
Moore G.W., Hutchins G.M. Coronary artery bypass grafts in 109 autopsied patients, statistical analysis of graft and anastomosis patency and regional myocardial injury. JAMA 1981;246:1785-1789.[Abstract/Free Full Text]
-
Weman S.M., Salminen U.S., Penttilä A., Männikkö A., Karhunen P.J. Post-mortem cast angiography in the diagnostics of graft complications in patients with fatal outcome following coronary artery bypass grafting (CABG). Int J Legal Med 1999;112:107-114.[Medline]
-
Ambrosio G., Becker L.C., Hutchins G.M., Weisman H.F., Weisfeldt M.L. Reduction in experimental infarct size by recombinant human superoxide dismutase. Circulation 1986;74:1424-1433.[Abstract/Free Full Text]
-
Ferreira R., Llesuy S., Milei J., et al. Assessment of myocardial oxidative stress in patients after myocardial revascularization. Am Heart J 1988;115:307-312.[Medline]
-
Seymour A.-M.L., Chatham J.C., Radda G.K. Effects of repeated low calcium perfusion on the rat heart. J Mol Cell Cardiol 1990;22:131-141.[Medline]
-
Lefer A.M. Role of selectins in myocardial ischemia-reperfusion injury. Ann Thorac Surg 1995;60:773-777.[Abstract/Free Full Text]
-
Nakanishi K., Zhao Z.-Q., Vinten-Johansen J., Lewis J.C., McGee D.S., Hammon J.W., Jr Coronary artery endothelial dysfunction after global ischemia, blood cardioplegia, and reperfusion. Ann Thorac Surg 1994;58:191-199.[Abstract]
-
Seccombe J.F., Schaff H.V. Coronary artery endothelial function after myocardial ischemia and reperfusion. Ann Thorac Surg 1995;60:778-788.[Abstract/Free Full Text]
-
Ikonomidis J.S., Rao V., Weisel R.D., Hayashida N., Shirai T. Myocardial protection for coronary bypass grafting. Ann Thorac Surg 1995;60:824-832.[Abstract/Free Full Text]
-
Buckberg G.D. Update on current techniques of myocardial protection. Ann Thorac Surg 1995;60:805-814.[Abstract/Free Full Text]
-
Mauney M.C., Kron I.L. The physiologic basis of warm cardioplegia. Ann Thorac Surg 1995;60:819-823.[Abstract/Free Full Text]
-
Jaquet L.M., Noirhomme P.H., Van Dyck M.J., et al. Randomized trial of intermittent antegrade warm blood versus cold crystalloid cardioplegia. Ann Thorac Surg 1999;67:471-477.[Abstract/Free Full Text]
-
Perrault L.P., Menasché P., Peynet J., et al. On-pump, beating-heart coronary artery operations in high-risk patients. An acceptable trade-off?. Ann Thorac Surg 1997;64:1368-1373.[Abstract/Free Full Text]
Accepted for publication April 6, 2000.
This article has been cited by other articles:
|
|
|
|
 
S. Grabherr, V. Djonov, K. Yen, M. J. Thali, and R. Dirnhofer
Postmortem Angiography: Review of Former and Current Methods
Am. J. Roentgenol.,
March 1, 2007;
188(3):
832 - 838.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
Y. Matsumoto, G. Watanabe, M. Endo, H. Sasaki, F. Kasashima, and I. Kosugi
Efficacy and safety of on-pump beating heart surgery for valvular disease
Ann. Thorac. Surg.,
September 1, 2002;
74(3):
678 - 683.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
