Death and adverse cardiac events after carotid endarterectomy

Transcription

1 Death and adverse cardiac events after carotid endarterectomy David J. Musser, MD, Gary G. Nicholas, MD, and James F. Reed III, PhD, Allentown, Pa. Purpose: This study evaluated operative mortality rate and adverse cardiac events after carotid endarterectomy. Efficacy of preoperative cardiac evaluation was studied and stroke mortality rate was determined. Methods: This was a retrospective review of 562 patients undergoing carotid endarterectomy at a 740-bed community hospital. Data were analyzed with X' analysis, logistic regression analysis, and Goldman criteria for cardiac risk. Results: The mortality rate was 1.6% (nine patients). There were 10 myocardial infarctions (1.8%). Six of these (1.1%) were fatal. The Goldman Index allowed us to classify 530 patients in a low-risk group (Goldman classes I and II, operative mortality rate;; 1.1%) and 32 patients in a high-risk group (Goldman classes III and IV, mortality rate ;; 9.4%). Independent risk variables were identified for myocardial infarction and overall operative death. These variables were then used to develop a probability model for prediction of operative death and adverse cardiac events. The stroke rate in the 562 patients was 0.7% (four patients). For the 345 patients with symptoms, the stroke rate was 0.6% (two patients); for the 217 symptom-free patients, it was 0.9% (two patients). The combined stroke mortality rate was 2.3%. For patients with symptoms, it was 2.9%; for symptom-free patients, it was 1.4%. Conclusions: Independent clinical variables can help determine patients at increased risk for perioperative myocardial infarction or operative death. Patients in Goldman classes III and IV are at increased risk for adverse events. Carotid surgery can be performed safely in our medical community. (J VAse SURG 1994;19: ) Stroke is the third leading cause of death in the United States. Because of the prevalence of the disease and the frequency with which therapeutic intervention is applied, careful assessment of all factors that may improve outcome should have a significant impact on the overall health of many patients with cerebral vascular disease. Hsia et al. 1 questioned the value of recently completed clinical trials documenting the efficacy of carotid endarterectomy in patients with symptoms because the mortality rate associated with carotid surgery in the From the Department of Surgery and the Research Department (Dr. Reed), Lehigh Valley Hospital, Allentown. Supported in part by research grant from the Dorothy Rider Pool Trust. Presented at the annual meeting of the Delaware Valley Vascular Society, September 20, 1992, Philadelphia, Pa. Reprint requests: Gary G. Nicholas, MD, Department of Surgery, Lehigh Valley Hospital, Cedar Crest and 1-78, PO Box 689, Allentown, PA Copyright 1994 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter /94/$ /1/48845 community at large exceeds that reported in these trials. Because myocardial infarction is the most common cause of death after carotid endarterectomy, preoperative cardiac evaluation, through clinical assessment or numerous qualitative and quantitative testing techniques, has been extensively studied. 2 Clinical evaluation has been used by numerous authors to identify constellations of risk factors that appear to place patients at high risk for adverse cardiac events after vascular surgical procedures. 3 In an earlier report of a series collected between 1967 to 1973 from the Cleveland Clinic, 4 the early mortality rate after carotid operation was 9.6% in patients with a history of coronary artery disease, compared with only 2.9% risk in patients with no clinical evidence of coronary artery disease. Eagle et al. 5 identified five clinical risk factors for coronary disease and reported a tenfold increased incidence of significant redistribution abnormalities on cardiac dipyridamole thallium stress testing and an increased number of adverse cardiac events in patients with these factors present. Despite these findings, the exact role of 615

2 616 Musser, Nicholas, and Reed JOURNAL OF VASCULAR SURGERY April 1994 extensive preoperative cardiac evaluation has been questioned in light of the low overall operative mortality rates achieved after careful clinical evaluation and selection of patients for operative intervention on that basis. 3 The purpose of this study was to determine the incidence of myocardial infarction, fatal myocardial infarction, and overall mortality rates among patients undergoing carotid endarterectomy. We evaluated this patient group for single factors or constellations of risk factors to determine those at increased risk for adverse coronary events and to justify invasive diagnostic procedures and aggressive therapeutic intervention, including trans luminal angioplasty or coronary artery bypass grafting. METHODS In the 5-year interval from January 1986 through December 1990, a total of 606 patients underwent carotid endarterectomy. The endarterectomy was done in combination with coronary artery bypass grafting in 44 patients. These patients were excluded from this consecutive series for risk analysis of carotid surgery. Patient records were reviewed in a retrospective fashion. Known risk factors for coronary artery disease, including those identified by both Eagle et al. 5 and Goldman et ai., 6 were tabulated for subsequent analysis. Every patient received preoperative evaluation consisting of a careful history, physical examination, and electrocardiography. Only 10 patients had dipyridamole thallium scans performed because this test was unavailable at our institution during most of the study. Peri operative screening with electrocardiogram and cardiac enzymes was at the discretion of the physician. After operation, 93% of the patients had an electrocardiogram at some time before discharge from the hospital. Myocardial infarction was defined as persistent electrocardiographic abnormalities with creatine kinase MB band levels greater than 5%. The 30-day mortality rate was recorded and all perioperative neurologic events were noted. Indwelling shunts were used in 89.9% of the operations and patch closure was used in 66.7% of the procedures. Perioperative characteristics were analyzed with a X 2 analysis for association. A stepwise logistic regression analysis was used to determine the probability of postoperative myocardial infarction and death on the basis of significant characteristics according to the following modef: e+ bx + cy +... p= e +bx+cy+... where P is the probability of a myocardial infarction; a is a constant; b, c,... are regression coefficients; and x, y,... are variables (factors) in the equation. The factors are coded according to the presence (x = 1) or absence (x = 0) of that factor. All selected variables were entered into a stepwise logistic regression model to determine the best set of independent predictors of myocardial infarction and of overall risk of death. Because of the large number of statistical tests, a Bonferroni-adjusted significance level of 0.01 was used to control for experimental error and to reduce the probability of detecting spurious differences. RESULTS During the 5-year interval from 1986 to 1990, carotid endarterectomies were performed in 562 patients at the Lehigh Valley Hospital. The mean age of the patients was 68 years, with a range of 40 to 90 years. Forty-percent of the patients were older than 70 years. There were 346 men (61.6%) and 216 women (38.4%). Of the 562 carotid endarterectomies, 345 (61.4%) were performed for hemispheric symptoms and 217 procedures (38.6%) were performed in symptom-free patients. All patients had carotid stenosis of at least 60% diameter demonstrated by angiography. Currently, all patients with asymptomatic stenosis are offered the option of enrolling in the Asymptomatic Carotid Atherosclerosis Study protocol. Patients who refuse to enter the protocol are currently considered for operation if they have a severe stenosis (> 80% diameter). Risk factors for atherosclerosis were tabulated for the 562 patients. Hypertension was present in 382, diabetes mellitus in 125, tobacco usage in 383, and hyperlipidemia in 479 (cholesterol > 260 mg/dl or triglycerides > 160 mg/dl). The overall mortality rate was 1.6% (nine patients). There were 10 (1.8%) myocardial infarctions in this series. Six (1.1 %) were fatal, accounting for 66.7% of the deaths. The incidence of adverse end points (myocardial infarction, fatal myocardial infarction, and death from all causes) in patients with clinical evidence of coronary artery disease (abnormal electrocardiogram, history of myocardial infarction, congestive heart failure, or angina pectoris) was evaluated. In the absence of any of these markers of coronary artery disease, the incidence of myocardial infarction was 0.7%. In the presence of one or more of these markers, the incidence of myocardial infarction was 2.1 %, but this was not significantly higher than in the group with none of these markers (p = 0.295).

3 JOURNAL OF VASCULAR SURGERY Volume 19, Number 4 Musser, Nicholas, and Reed 617 Table I. Risk factors for myocardial infarction in 562 patients undergoing carotid endarterectomy Factor Patients MJ* % p Atrial fibrillation Previous MI Emergency operation Stable angina Premature ventricular contractions on ECG Intraoperative hypotension t Intraoperative dopamine History end-stage renal disease I MI, Myocardial infarction; EeG, electrocardiogram. *Total number of myocardial infarctions was 10. tsystolic pressure < 100 mrn Hg. Similarly, there was no significant difference between the two groups when the end points of fatal myocardial infarction and death from all causes were analyzed (p = and p = 0.361, respectively). Preoperative risk factors and intraoperative events were examined for association with perioperative myocardial infarction and death. Significant relationships between independent risk variables and the development of myocardial infarction are tabulated in Table I. A stepwise logistic regression analysis indicated factors that were predictive of the development of perioperative myocardial infarction and allowed the derivation of the probability equation. Significant predictive factors were a history of endstage renal disease (hemodialysis or peritoneal dialysis), premature ventricular contractions on preoperative electrocardiogram, and stable angina (Table II). Analysis of factors related to the overall mortality rate in the series of 562 procedures revealed five individual factors associated with operative death (Table III). A stepwise logistic regression analysis with these factors found three to be significantly predictive of operative death (Table IV). Emergency operation was performed in 11 patients who had crescendo transient ischemic attacks or in whom angiography revealed a preocclusive stenosis or unstable thrombus. Patients were classified according to the original criteria of Goldman et al. 6 Of the 562 patients, 81 % (455) were in class I, 13.3% (75) were in class II, 4.6% (26) were in class III, and 1.1% (6) were in class IV. The incidence of myocardial infarction, fatal myocardial infarction, and death from all causes increased with increasing Goldman class assignment. There were only six patients in class IV, and no adverse end points were noted. For the Goldman classes I, II, and III, the rates of myocardial infarction were 0.9%, 2.7%, and 15-4%, respectively. The rates of fatal myocardial infarction were 0-4%, 1.3%, and Table II. Equation for estimation of myocardial infarction risk: Logistic regression of significant factors in 562 patients undergoing carotid endarterectomy exp(c1xi + CzXz + C 3X 3 + C~ + C) P MI = "..--=-'---,-=-=., ,~,------,,:-=---=-=--=- 1 + exp(c1xl + CzXz + C 3 X 3 + C 4 J4 +C) C I = C 2 = C3 = C 4 = C = Xl = Stable angina X 2 = Premature ventriculat contraction X3 = Intraopertive hypotension X 4 = End-stage renal disease Sample: A patient with premature ventricular contractions and end-stage renal disease: exp( ) PM! = 1 + exp( ) PM! = /( = = 68% Intraoperative hypotension defined as systolic pressure < 100 mrn Hg for 5 minutes or more. Premature ventricular contractions noted on preoperative electrocardiogram. End-stage renal disease defined as hemodialysis or peritoneal dialysis. 11.5%, respectively. The overall mortality rates were 0.7%,4.0%, and 11.5%, respectively. In Table V, comparison of the patients in Goldman classes I and II with the patients in the Goldman classes III and IV is presented. This comparison revealed a highly significant difference between these groups with regard to adverse cardiac events and overall operative mortality rate. There were 102 patients (18.1 %) in this series who underwent coronary artery bypass grafting at some time before their carotid endarterectomy. These were patients with either symptomatic coronary artery disease or significant coronary occlusive disease noted on cardiac catheterization after a positive stress test. They were initially referred to our

4 618 Musser) Nicholas) and Reed JOURNAL OF VASCULAR SURGERY April 1994 Table III. Risk factors for operative death in 562 patients undergoing carotid endarterectomy Factor Patients Deaths* % p Atrial fibrillation 34 Previous myocardial infarction Emergency operation Premature ventricular contractions on ECG Intraoperative hypotensiont History of end-stage renal disease EGG) Electrocardiogram. *Total number of deaths was nine. tsystolic pressure < 100 mm Hg. Table IV. Equation for estimation of mortality risk: Logistic regression of significant factors in 562 patients undergoing carotid endarterectomy exp(cjxj + C 2 X 2 + C3X3 + C) PM ali =.,--,...:-,---;~c;--.,...::.;~,...,""::""'''::'';';'--,--,-=" orr ry 1 + exp(cjxj + C 2 X 2 + C 3 X 3 + C) C j = C 2 = C 3 = C = Xl = Emergency operation X 2 = Premature ventricular contraction X3 = End-stage renal disease Sample: A patient with premature ventricular contractions and end stage renal disease: exp( ) PMorraliry = 1 + exp( ) PMortality = /( ) = See Table II for definitions. institution for evaluation of their coronary artery disease. The incidence of myocardial infarction in this group was 1.0% (1/102) after carotid operation. Cardiac catheterization was done as part of the preoperative evaluation in 26 patients. In 22 of these patients, abnormalities were found but were not considered severe enough to warrant interventional therapy. The patients were selected for catheterization on the basis of symptoms and positive results of an exercise thallium stress test. The abnormalities consisted of stenotic lesions of less than 50% distributed throughout the right and left coronary systems. None of these 22 patients had severe triple-vessel disease and none had left main or equivalent disease. In none of these patients was there. an adverse cardiac event or operative death after their carotid operation. In this series of 562 patients, the perioperative stroke rate was 0.7% (four patients) and the mortality rate was 1.6% (nine patients), for a combined stroke mortality rate of 2.3%. For the 345 patients with symptoms, the stroke rate was 0.6% (two patients) and the mortality rate was 2.3% (eight patients), for a combined stroke mortality rate of2.9%. In the 217 symptom-free patients, the perioperative stroke rate was 0.9% (two patients) and the mortality rate was 0.5% (one patient), for a combined stroke mortality rate of 1.4%. DISCUSSION The purpose of this study was to determine the incidence of myocardial infarction, fatal myocardial infarction, and death from all causes for patients undergoing carotid endarterectomy at our institution. We evaluated the results of 562 patients undergoing consecutive carotid endarterectomies in a 5-year interval to determine clinical factors and constellations of factors that would define patients at high risk for adverse cardiac events after carotid operation. We also evaluated our major morbidity, stroke, after carotid endarterectomy. For patients undergoing carotid endarterectomy, the overall mortality rate was 1.6%. These results were similar to those reported by other authors and are well within the guidelines recommended by the American Heart Association for institutions and surgeons performing carotid endarterectomy. A combined stroke mortality rate of less than 3% for patients undergoing carotid surgery for asymptomatic stenosis was recommended. 8 In a series collected between 1967 and 1973, Hertzer and Lees 4 reported a patient mortality rate of 3% and a procedural mortality rate of2.6% after carotid surgery. In 1989 the mortality rate for carotid endarterectomy in Medicare beneficiaries was 2.5%.1 Adverse cardiac events in our series of patients occurred after 10 of the 562 operative procedures. The myocardial infarction rate was 1.8% and the fatal myocardial infarction rate was 1.1%. Taylor et al. 3 reported an overall operative mortality rate of 2.2% in 491 patients undergoing 534 vascular reconstruc-

5 JOURNAL OF VASCULAR SURGERY Volume 19, Number 4 Musser, Nicholas, and Reed 619 Table V. Goldman risk index for adverse cardiac events and death from all causes in 562 patients undergoing carotid endarterectomy MI(%) FatalMI (%) Overall mortality rate Classes No. No. % I and II III and IV P Value No. % No. % tions. They noted a myocardial infarction rate of 3.9%, with fatal myocardial infarctions occurring in 0.75%. Only three patients in their series underwent preoperative coronary artery bypass grafting. In Hertzer and Lees' series of 390 carotid endarterectomies/ they noted a 1. 79% myocardial infarction rate and a 1.54% fatal myocardial infarction rate in their group of patients operated on between 1967 and Only five of the patients in that series underwent combined coronary artery bypass grafting and carotid endarterectomy. Similar results were reported by Riles et al. 9 in their series of 683 carotid endarterectomies. They noted a fatal myocardial ifarction rate of 0.73% and an overall myocardial infarction rate of 2.34%. They determined that patients receiving intraoperative pressor therapy were at markedly increased risk for adverse cardiac events. More recently, Eagle et al. 5 reported a myocardial infarction rate of 4.5% after major vascular operations and a fatal myocardial infarction rate of3.0% in their series of 200 patients. The asymptomatic carotid endarterectomy trial conducted at the Mayo Clinic lo was terminated early because of the increased frequency of myocardial infarction in the surgical group. This was attributed to the absence of the use of aspirin in patients in that arm of the study. Because myocardial infarction is the most common cause of death after carotid endarterectomy, the definition of patients at increased risk for death after carotid surgery has been of great interest to vascular surgeons and cardiologists. 2 Coronary artery disease has been found in as many as 65% of patients evaluated before vascular reconstructive operations. 11 Even in the absence of symptoms of cardiac disease, silent coronary artery disease has been noted in 25 % of patients evaluated before vascular surgical intervention. 12 The low overall operative mortality rate in our series and those reported by others mean that any cardiac intervention without an effort to select patients at high risk may be unable to improve the early outcome when the operative mortality and morbidity rate from cardiac catheterization and intervention, with either transluminal angioplasty or coronary bypass grafting, is superimposed sequentially on this patient group. We therefore looked at clinical factors in this group of 562 patients to determine characteristics that would allow selection of patients at high risk for adverse cardiac events after carotid operation. Despite the fact that the use of tobacco is a totally avoidable risk factor, it was the most common risk noted in this series of patients. In the 135 patients who did not exhibit any of the common characteristics of coronary artery disease (abnormal electrocardiogram, previous myocardial infarction, angina, congestive heart failure), the incidence of myocardial infarction was 0.7%. In the 427 patients who had one or more of these markers, the incidence of myocardial infarction was 2.1%. These two groups were not statistically distinct. It thus appears that in our series these simple, easily obtainable clinical characteristics did not allow definition of a subset of patients at increased risk for adverse cardiac events after carotid operation. We analyzed our patient data further through univariate analysis of risk factors that have been noted by many other authors. These results are tabulated in Table I. In our series, atrial fibrillation was noted to be associated with a 10.5% incidence of perioperative myocardial infarction. This factor has not been reported by previous authors assessing risk factors for perioperative myocardial infarction after vascular reconstructive operations. By means of logistic regression analysis, four factors were found to be statistically associated with an increased risk of myocardial infarction (p < 0.01). This analysis allowed development of the equation presented in Table II, which provides an estimate of the probability of myocardial infarction after carotid operation. The clinical factors we analyzed were evaluated with respect to overall operative mortality rate and are tabulated in Table III. These data were analyzed through logistic regression analysis and the resulting significant factors (p < 0.01) allowed development of the equation for the predictive model for operative death after carotid endarterectomy (Table IV).

6 620 Musser) Nicholas) and Reed JOURNAL OF VASCULAR SURGERY April 1994 We tabulated the four Goldman classes for risk of cardiac complications in our patients series. Through three of the four classes, the perioperative myocardial infarction and fatal myocardial infarction rates increased as the number of risk factors defined by the Goldman Index increased. There were only six patients in class IV and these had no adverse end-point events. In Table V, we contrast the perioperative risks of adverse cardiac events and death of classes I and II, the lower risk groups, to those of the higher-risk groups in classes III and IV. In the 530 patients in classes I and II, the overall operative mortality rate was only 1.1 %, with a myocardial infarction rate of 1.1 %. In contrast, of the 32 patients in Goldman classes III and IV, the overall mortality rate was 9.4%, the myocardial infarction rate was 12.5%, and the fatal myocardial infarction rate was 9.4%. The difference between these two groups was statistically significant for mortality rate and adverse cardiac events. In the report from the Cleveland Clinic,4 the early mortality rate was 9.6% in patients with clinically identifiable coronary artery disease, whereas it was only 2.9% in those without clinical coronary artery disease. Eagle et al. 5 noted five clinical predictors of postoperative ischemic cardiac events after vascular surgery: Q waves, history of ventricular ectopic activity, diabetes, advanced age (> 70 years), and angina pectoris. Data reported by others, however, have not shown clinical parameters to be helpful in stratification of patients with regard to risk of operative death and adverse cardiac events. Cutler et al. 13 noted that dipyridamole thallium screening before vascular reconstruction in 262 patients was the only significant predictor of early and late adverse cardiac events in their patient group. Only the addition of congestive heart failure to the findings of the fixed deficit on dipyridamole thallium scanning increased the reliability of their predictive model. Similarly, Lette et al. 14 found that clinical parameters did not predict outcome after major vascular reconstruction in their 125 patients; however, dipyridamole thallium scanning was useful in stratification of these patients regarding perioperative risk for adverse cardiac events. Eagle et al. 5 used clinical criteria to stratify patients as to degree of risk for adverse cardiac events after vascular surgical procedures. They used five clinical predictors to define the need for dipyridamole thallium screening and were able to determine three groups of patients at risk for adverse cardiac events. Those patients at intermediate risk with thallium redistribution had a 29.6% incidence of adverse cardiac events, as compared with an incidence of 3.2% adverse events in those patients who did not demonstrate thallium redistribution. According to logistic regression analysis, the probability equations for perioperative myocardial infarction or operative death are tabulated in Tables II and IV. These equations may be useful in predicting untoward events after carotid endarterectomy. The clinician may use these probabilities in assessing the need for in-depth preoperative cardiac evaluation. The validity of these probability models requires prospective evaluation in a large clinical series. For the 562 patients undergoing carotid endarterectomy, the stroke mortality rate was 2.3%. The overall stroke rate was 0.7% (four patients). For the 345 patients with symptoms, the stroke rate was 0.6% (two patients). These results are similar to those published by the major multiinstitutional trials for patients with symptoms undergoing carotid operations. In the North American Symptomatic Carotid Endarterectomy Trial,15 the combined stroke mortality rate was 2.7%, with major strokes occurring in 2.1 % of the 328 patients with symptoms who were randomly assigned to the surgical group. In the European trial,16 the stroke mortality rate for the 455 patients with severe stenosis randomly assigned to the surgical group was 2.86%. Individual institutions have reported similar results. Maini et al. 17 noted a 2.4% stroke rate in their series of215 patients and De Bord et al. 1S found a 1.2% stroke rate in the 303 patients who underwent carotid endarterectomy at their institution during a 5-year period. The results from our institution are similar to those reported in multiinstitutional international studies, as well as the results achieved at individual institutions in the community and university settings. The data presented here are limited by the retrospective nature of the study and the fact that the diagnosis of myocardial infarction was not standardized. At the discretion of the clinicians, a number of the patients had routine peri operative screening with electrocardiograms and cardiac enzyme assays, although this was not uniform throughout the study. Other patients were evaluated for myocardial infarction only on the basis of clinical symptoms or signs. Although this undoubtedly led to underreporting of myocardial infarction, the end points of fatal myocardial infarction and death from all causes are clearly defined and support our conclusions from the data. The patients reported in this series predate the availability of adenosine thallium stress testing in our community. This lack did, however, allow us to evaluate results of selection of patients on the basis of clinical status. The results must also be viewed with

7 JOURNAL OF VASCULAR SURGERY Volume 19, Number 4 Musser) Nicholas) and Reed 621 the knowledge that 102 patients with coronary disease had bypass operations before carotid endarterectomy. These patients are included in this consecutive series because they represent a segment of the population referred to our facility for carotid surgery. The use of the Goldman Risk Index was originally described in the evaluation of patients undergoing general surgical, urologic, and orthopedic procedures. 6 The prevalence of coronary artery disease is high in patients being evaluated for vascular reconstructive surgery and may have differed from that in the population originally evaluated by Goldman et al. 6 Although this difference could adversely affect the use of this Index, we found a good correlation. The incidence of silent coronary artery disease, 25% as reported by Urbinati et al.,12 would indicate that, despite the fact that these patients are overlooked on the basis of clinical evaluation, the short-term results of perioperative mortality rates in our series were not affected adversely. As documented by Cutler et al.,13 however, the long-term survival after carotid operation appears to be significantly affected by abnormal results of a cardiac evaluation study. Our data do not address the long-term results of carotid operation or the possible benefit that may be provided by coronary artery bypass grafting in certain subsets, such as those with left main coronary disease, three-vessel disease, and impaired ventricular function. 19 Our results agree with those published by Taylor et al., 3 indicating that intensive cardiac evaluation and intervention cannot be justified in light of the low operative mortality rate achieved with most vascular reconstructive procedures in patients who do not demonstrate significant clinical risk factors. The short-term benefit of invasive screening and therapeutic intervention may be limited by the morbidity and mortality associated with the procedures. Because of the 1.1 % overall mortality rate after carotid endarterectomy in the group with low cardiac risk (Goldman classes I and II), it is unlikely that cardiac intervention with either percutaneous transluminal coronary angioplasty or coronary artery bypass grafting will be able to improve the overall early results in this group. It does appear, however, that patients identified as being at increased risk on the basis of clinical criteria (Goldman classes III and IV) should benefit from intervention through further diagnostic and therapeutic procedures. In our series, the operative mortality rate for the 102 patients who had coronary artery bypass grafting at some time before carotid endarterectomy was only 1.0% after the carotid operation. The use of adenosine thallium stress testing in this high-risk group may identify those patients most likely to be benefited. It is now our policy to perform adenosine thallium testing before operation on patients in all Goldman classes III and IV before recommending carotid endarterectomy. Patients with symptomatic carotid disease are considered for combined coronary grafting and carotid operation if cardiac catheterization identifies severe disease. Alternatively, medical therapy is considered if the operative risks are considered prohibitive. Analysis of cardiac risk factors should allow selection of patients who may benefit from preoperative cardiac screening and intensive perioperative hemodynamic management, minimizing unnecessary testing in the patients at lower risk. REFERENCES 1. Hsia DC, Krushat WM, Moscoe LM. Epidemiology of carotid endarterectomies among Medicare beneficiaries. J VASC SURG 1992;16: Yeager RA, Moneta GL, McConnell DB, N euwelt EA, Taylor LM Jr, Porter JM. Analysis of risk factors for myocardial infarction following carotid endarterectomy. Arch Surg 1989; 124: Taylor LM Jr, Yeager RA, Moneta GL, McConnell DB, Porter JM. The incidence of peri operative myocardial infarction in general vascular surgery. JVASC SURG 1991;15: Hertzer NR, Lees CD. Fatal myocardial infarction following carotid endarterectomy. Ann Surg 1981;194: Eagle KA, Coley CM, Newell JB, et al. Combining clinical and thallium data optimizes preoperative assessment of cardiac risk before major vascular surgery. Ann Intern Med 1989;110: Goldman L, Caldera DL, Nussbaum SR, et al. Multifactorial index of cardiac risk in noncardiac surgical procedures. N Engl J Med 1977;297: Cox DR. The analysis of binary data. New York: Champion & Hall, Beebe HG, Clagett GP, DeWeese JA, et al. Assessing risk associated with carotid endarterectomy: a statement for health professionals by an ad hoc committee on carotid surgery standards of the Sttoke Council, American Heart Association. Stroke 1989;20: Riles TS, Kopelman I, Imparato AM. Myocardial infarction following carotid endarterectomy: a review of 683 operations. Surgery 1979;85: Mayo Asymptomatic Carotid Endarterectomy Study Group. Results of a randomized controlled trial of carotid endarterectomy for asymptomatic carotid stenosis. Mayo Clin Proc 1992;67: Hertzer NR, Beven EG, Young JR, et al. Coronary artery disease in peripheral vascular patients: a classification of 1000 coronary angiograms and results of surgical management. Ann Surg 1984;199: Urbinati S, Di Pasquale G, Andreoli A, et al. Frequency and prognostic significance of silent coronary artery disease in patients with cerebral ischemia undergoing carotid endarterectomy. Am J CardioI1992;69: Cutler BS, Hendel RC, Leppo JA. Dipyridamole-thallium

8 622 Musser) Nicholas) and Reed JOURNAL OF VASCULAR SURGERY April 1994 scintigraphy predicts perioperative and long-term survival after major vascular surgery. J VAse SURG 1992;15: Lette J, Waters D, Lassonde J, et al. Multivariate clinical models and quantitative dipyridamole-thallium imaging to predict cardiac morbidity and death after vascular reconstruction. J VAse SURG 1991;14: North American Symptomatic Carotid Endatterectomy Trial Collaborators. Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade stenosis. N Engl J Med 1991;325: European Carotid Surgery Trialists' Collaborative Group. MRC European Carotid Surgery Trial: interim results for symptomatic patients with severe (70-99%) or with mild (0-29%) carotid stenosis. Lancet 1991;337: Maini BS, Mullins TF III, Catlin J, O'Mara P. Carotid endarterectomy: a ten-year analysis of outcome and cost of treatment. J VAse SURG 1990;12: De Bord JR, Marshall WH, Wyffels PL, Marshall JS, Humphrey P. Carotid endarterectomy in a community hospital surgical practice. Am Surg 1991;57: CASS principle investigators and their associates. Myocardial infarction and mortality in the Coronary Artery Surgery Study(CASS) randomized trial. N Engl J Med 1984;310: Submitted March 22, 1993; accepted May 21, 1993.