ery: Comparison of Predicted and Observed Resu ts

Preoperative Risk Assessment in Cardiac Sur K ery: Comparison of Predicted and Observed Resu ts Forrest L. Junod, M.D., Bradley J. Harlan, M.D., Janie Payne, R.N., Edward A. Smeloff, M.D., George E. Miller, Jr., M.D., Paul B. Kelly, Jr., M.D., Kenneth A. Ross, M.D., Kuppe G. Shankar, M.D., and John P. McDermott, M.D. ABSTRACT In the present climate of quality-assurance policies, rigorous requirements for informed consent, and a constantly changing patient population, a system of preoperative risk assignment and postoperative correlation was developed to monitor and evaluate surgical performance. Patients were categorized by operation, priority (emergent, urgent, elective), New York Heart Association Functional Class, and risk. Risk was assigned before operation using data from the Coronary Artery Surgery Study (CASS) and the recent literature. Data were collected by a full-time data manager and were stored and analyzed by computer. From January 1, 1984, to July 1, 1985, 1,33 patients underwent operation for acquired disease. This group included 913 patients undergoing isolated primary coronary artery bypass grafting (CABG). The comparison of predicted and observed results showed Predicted Operative Mortality <2% 2-5% 6-1% 11-5% >5% Observed Operative Mortality, All Patients.6% (3/516).6% (3/486) 5.2% (81155) 1.5% (1/95) 54.9% (28/51) Observed Operative Mortality, Isolated Primary CABG.6% (3/59).3% (1/323) 8.5% (4/47) 9.1% (2/22) 33.32 (4/12) For patients undergoing isolated primary CABG, the elective group had an operative mortality of.6% (2/329); the urgent group, 1.1% (5/45); and the emergent group, 5.2% (7/134). Preoperative risk assignment is an effective method of quality assurance. Female sex and age older than 6 years, which predicted an operative mortality of 2 to 5% in the CASS study and other recent series, did not predict a similar risk in our series. Risk assessment is becoming increasingly important in cardiac surgery. Patients desire an accurate statement of From the Division of Cardiovascular Surgery, Sutter Memorial Hospital, Sacramento, CA. Presented at the Twenty-second Annual Meeting of The Society of Thoracic Surgeons, Washington, DC, Jan 27-29. 1986. Address reprint requests to Dr. Junod, 531 F St, Suite 312, Sacramento, CA 95819. the risk they are to assume for the treatment of their disease. An informed consent is a medicolegal necessity. Quality-assurance programs are growing, thereby increasing the scrutiny of cardiac surgery and the need for accurate data relating to the performance of cardiac operations. Changing methods of surgical management are probably altering risk and consequently making it important to define how risks are being changed and to identify areas for further improvement. For these reasons, we initiated a system of preoperative risk assessment in 1984. This study reports the results of that system over a period of eighteen months. Material and Methods From January 1, 1984, to July 1, 1985, 1,33 patients had preoperative assignment of risk before undergoing operation for acquired heart disease. Distribution by operation and predicted operative mortality are shown in Table 1. Patients were assigned by risk of operative death into one of five categories: less than 2%, 2 to 5%, 6 to lo%, 11 to 5%, and higher than 5%. The patients and their families were quoted the same risk, and the factors determining their risk were explained. Risk assignment was made using data from the Coronary Artery Surgery Study (CASS) and the recent literature [ 1-141. Surgical priority was assigned using definitions similar to those in the CASS study [l]. Emergent status (15%) required operation within 24 hours of catheterization or within 24 hours of a major change in clinical condition. Urgent status (44%) required operation during the same hospitalization and within four days of catheterization. Elective status (41%) included all other patients. Most patients assigned a risk of less than 2% were those undergoing primary isolated coronary artery bypass grafting (CABG) with normal or nearly normal ventricular function and no serious associated disease. Patients assigned to risk 2 to 5% had isolated valve replacement in the presence of good ventricular function, repeat CABG, or primary isolated CABG. The variables that were used to assign a risk of 2 to 5% for isolated primary CABG were age of 6 years or older (223 patients, 69.%), female sex (113 patients, 35.%), ejection fraction of less than 4% (44 patients, 14.%), acute myocardial infarction within fourteen days (21 patients, 6.5%), nonelective priority (249 patients, 77.%), and preoperative intraaortic balloon pumping (21 patients, 6.5%). Most patients assigned a risk of 6 to 1% had isolated 59 Ann Thorac Surg 43:59-64, Jan 1987

6 The Annals of Thoracic Surgery Vol 43 No 1 January 1987 Table I. Distribution by Operation and Predicted Operative Mortality" Predicted Operative Mortality Operation <2% 2-57" 6-1'5, 11-5% >5% Isolated primary CABG, 913 Isolated redo CABG, 98 Isolated AVR, 67 AVR + CABG, 61 Isolated MVR, 4 MVR + CABG, 42 AVR + MVR 2 CABG, 25 Other, 57 323 (35) 7 (71) 36 (54) 16 (26) 23 (58) 3 (7) 2 (8) 13 (23) 47 (5) 14 (14) 22 (33) 32 (53) 1 (25) 18 (43) 3 (12) 9 (16) 22 (2) 9 (9) 7 (1) 13 (21) 4 (1) 11 (26) 12 (48) 17 (3) "Numbers in parentheses are percents within each operation group. CABG = coronary artery bypass grafting; AVR = aortic valve replacement; MVR = mitral valve replacement valve replacement with poor ventricular function, valve replacement with CABG, and CABG in the presence of serious complicating features, such as emergent operation, preoperative requirement of intraaortic balloon pump support, or severely depressed left ventricular function. Patients assigned a risk of 11 to 5% usually had multiple procedures or several complicating conditions. Most patients assigned a risk of greater than 5% were in cardiogenic shock and were certain to die without surgical intervention. Data were collected by a full-time data manager and were stored and analyzed by computer. Statistical comparison of groups was made by chi-square analysis with statistical significance being any p value less than.5 (Gary Grunkemeier, Ph.D., Medical Data Research, Portland, OR). Primary isolated CABG was performed in 913 patients. New York Heart Association (NYHA) Functional Classes 111 and IV accounted for 75.8% of the group. Only 6 of the patients were in Functional Class I. The age range was 33 to 88 years with a mean of 62.6 years. From 1 to 9 coronary anastomoses were performed per patient with a mean of 4.4. Technique of Operation All operations were performed using cardiopulmonary bypass and almost all, with moderate systemic hypothermia. Short-acting narcotic anesthesia was used. Myocardial preservation was achieved with a hypothermic hyperkalemic crystalloid solution containing lidocaine hydrochloride or blood cardioplegia. Blood cardioplegia usually was used after initial crystalloid arrest and cooling for patients with long cross-clamp times (greater than ninety minutes), left ventricular hypertrophy, cardiogenic shock, and acutely occluded coronary arteries. Myocardial temperature was usually monitored and kept lower than 2 C. Distal coronary anastomoses were performed first in the arrested heart with venting through the aorta. Aortic anastomoses were performed over a partial occluding clamp in the beating, nonvented heart. Internal mam- mary arteries were used in 2% (183/913) of all patients and in 41% (4/98) of patients undergoing repeat CABG. Arm veins were used in 6 patients for primary operations and in 6 for a repeat operation. Results The observed mortality compared with the predicted mortality is shown in Table 2. The observed mortality matches the predicted mortality in all risk groups except the 2 to 5% group where the observed risk was lower than the predicted risk. Therefore, patients were not given an erroneously low risk. Operative deaths by surgical priority demonstrated a statistically higher risk in the emergent priority than the elective priority. In the elective group, 11 (2.1%) of all 533 patients so classified died and 2 (.6%) of the 329 patients having isolated primary CABG. In the urgent group, 15 (2.6%) of all 58 patients so categorized died and 5 (1.1%) of the 45 having isolated primary CABG. In the emergent group, 26 (13.7%) of 19 patients died and 7 (5.2%) of the 134 patients having isolated primary CABG (p <.1 compared with elective and urgent groups). Operative mortality by NYHA Functional Class is shown in Table 3. Table 4 shows operative mortality by age and sex. Overall, there was no increased risk associated with increased age. There was 1 operative death in 188 women undergoing nonemergent primary isolated CABG. The only subset of patients with higher risk was women more than 7 years old. Overall, there was no difference in risk between men and women. Table 5 shows complications and causes of death in patients undergoing isolated primary CABG. Comment Risks of operation for cardiac disease have decreased over the years. Many technical advances have contributed to the reduction in risk, including myocardial preservation, advances in microvascular surgery and perioperative care, and improved diagnostic and monitoring techniques. Indications for operation have

~~ ~~~~~~~ ~~ ~ 61 Junod, Harlan, Payne, et al: Preoperative Risk Assessment in Cardiac Surgery Table 2. Predicted and Observed Mortality by Operation" Risk Groups Operation <21 2-5% 6-1%) 11-51 >5'% Totals Isolated primary CABG Isolated redo CABG Isolated AVR AVR + CABG Isolated MVR MVR + CABG AVR + MVR 2 CABG Other 3159 (.6) 11 /6 1/323 (.3) 4/47 (8.5) 2/7 (2.9) 2/14 (14.3) 136 1/22 (4.5) /16 132 123 1/1 13 118 /2 13 /13 19 2/22 (9.1) 219 (22.2) 17 2/13 (15.4) 14 2/11 (18.2) 1/12 (8.3) 1117 (5.9) 4/12 (33.3) 2/4 (5.) 112 (5.) 2/3 (66.7) 7/1 (7) 5/8 (62.5) 7/12 (58.3) 14/913 (1.5) 8/98 (8.2) 2/67 (3.) 2/61 (3.3) 3/4 (7.5) 9/42 (21.4) 6/25 (24) 8/57 (14) Total 31516 (.6) 3/486 (.6) 8/155 (5.2) 1195 (1.5) 28151 (54.9) 52/1,33 (4.) "Numbers in parentheses are percents. CABG = coronary artery bypass grafting; AVR = aortic valve replacement; MVR = mitral valve replacement. evolved as both medical and surgical treatments have improved. The need to inform patients requires assessment of risks prior to treatment. Risks of operative death for valve replacement have shown overall improvement. In 1984, Cohn and associates [15] reported a drop in mortality for isolated aortic valve replacement (AVR) from 4.5 to 1.%. Nunley and co-workers [12] thought the risk of CABG with AVR was decreased because of reduced cardiac deaths (approximately 5%) in 1981. In a 1983 report, Lytle and his colleagues [13] suggested a higher risk in women, older patients, and those with multivessel disease. That report covered a fifteen-year experience in which the operative mortality for AVR with revascularization improved from 8 to 3%. The report further suggested no significant difference between isolated AVR and AVR plus CABG. Morbidity, particularly perioperative myocardial infarction, greatly increased the mortality. Our study supports the conclusion that there is little difference in the risk for AVR alone or AVR with CABG (3.% versus 3.3%). Miller and associates [lo] reported a wide range (5 to 56%) of operative mortality for different subsets of patients having mitral valve replacement. The variation depended on etiological factors as well as the lesion (steno- sis or regurgitation). The addition of revascularization or ventricular resection after infarction greatly increased the observed mortality. A report by Lytle and coworkers [16] discussed factors in hospital mortality for mitral valve replacement and CABG in a series with an operative mortality of approximately 7%. Preoperative cardiac enlargement, paced rhythm, atrial fibrillation, left main coronary artery stenosis, and increased bilirubin level were identified as increasing mortality. Scott and colleagues [17] found New York Heart Association Functional Class, previous myocardial infarction, and hepatic dysfunction to be powerful determinants of surgical outcome in mitral valve disease. From our series, we believe mitral valve replacement with or without revascularization can be achieved with low risk. One of 54 patients died in the groups with a predicted mortality of 1% or less. Those patients having emergency operation, massive regurgitation, low ejection fractions (less than 4%), and complex multisystem problems had an extremely high operative risk. Eleven of 28 patients died, an operative mortality of 39.3%. Magovern and coworkers [18] reported extreme risk (45%) for mitral valve replacement when the patients required emergency operation for cardiogenic shock. Functional class has been shown by Magovern [MI, Table 3. Operative Mortality by New York Heart Association Functional Cla~s~,~ Patients with Isolated Class All Patients Primary CABG I 113 16 I1 1279 /215 I11 8156 (1.6) 4/334 (1.2) IV W55 (8.7) ( p <.1) 11358 (2.8) (p <.2) "Numbers in parentheses are percents bthe p values are for Class IV compared with Class II Table 4. Operative Mortality by Age and Sex for Patients Having Isolated Primary CABG" Age (yr) All Patients Male Female < 5 184 /7 /14 5-59 Y249 (.8) 2213 (.9) /36 6-69 7/348 (2.) NS 6/25 (2.4) NS 1/98 (1.) 2 7 5/232 (2.2) NS 11148 (.7) 4/84 (4.8) p <.5 Total 14/913 (1.5) 9/681 (1.3) 51232 (2.2) NS 'Numbers in parentheses are percents. CABG = coronary artery bypass grafting; NS = not significant to p <.5.

62 The Annals of Thoracic Surgery Vol 43 No 1 January 1987 Table 5. Complications in Patients Having Isolated Primary CABG No. of Operative Complication Patients Deaths Cardiac (53 patientd65 events) Periop MI 26 2 Recatheterization within thirty days of operation 26 Cardiac arrest 11 4 Pacer for heart block 2 Central nervous system 14 2 Pulmonary 16 3 Pulmonary embolus 15 Tracheostomy 1 Postop bleeding requiring reoperation 14 2 Other 15 1 CABG = coronary artery bypass grafting; MI = myocardial infarction; CVA = cerebrovascular accident. Scott [17, 191, and their associates to be significant in determining operative result. Patients in NYHA Class IV had a mortality of 26% with mitral valve replacement [18]. Our series of operations other than CABG for 97 Class IV patients resulted in 28 deaths (28.9% operative mortality). Repeat operation added to the risk of CABG or AVR in this study. Repeat CABG with a less than 1% predicted operative mortality had a 4.7% operative mortality (4/85) compared with.9% (8/879) for primary CABG. For groups with a risk of greater than lo%, the mortality doubled for reoperations. With AVR, 1 death occurred in 115 primary operations, whether isolated or with CABG (.9% operative mortality). For redo AVR with or without CABG, 3 of 17 patients died, an operative mortality of 17.6%. The CASS report by Myers and associates [2] for isolated primary CABG had an overall mortality of 2.4%. Left main stenosis increased the risk to 3.8%. Threevessel disease increased the risk of operation. Poor left ventricular function showed a higher mortality [21]. Differences in survival curves for CASS patients at three months were related to differences in surgical mortality. Within the fifteen institutions of the CASS report, the operative mortality ranged from.3 to 6.6% [l], a finding that raises the possibility that the overall survival results were adversely affected by high operative mortality. The CASS researchers concluded that age greater than 6 years and female sex (or height) affected operative mortality [3]. Surgical priority influenced the outcome as well. Our study and other recent studies suggest that contemporary surgical results are superior to the results reported in the CASS literature. Miller and co-workers [6] presented a series in which sex and age were neutralized as factors for the patients undergoing isolated CABG. Cosgrove and associates [22] reported excellent results and drew the conclusion that age and sex persisted as risk factors in 198 through 1982. As shown in Tables 6 and 7, our experience is similar to that of the Cleveland Clinic. However, our data support the decreased importance of age as a determinant in the seventh and eighth decades. Only women showed a statistically significant difference in patients older than 7 years. In our study, most of the deaths (4 of 5) occurred in women undergoing emergency operation for unremitting angina; therefore, factors other than sex or age would appear to account for the difference. With continued improvement in microvascular techniques, perhaps small vessels will not represent the challenge previously considered to be a mortality factor in women. Improvements in preoperative care have enhanced the survival of those more than 7 years old, as reported by Montague and associates [23]. For the surgical population reported here, emergency surgical priority had a highly significantly different risk from elective priority (p <.1). Cosgrove and colleagues [22] also reported a sevenfold increase in risk for patients having emergency coronary bypass operations compared with those operated on electively. Congestive heart failure replaced emergency priority as the number one determining factor in the later patients in their series. Comparison of the four series must be cautious for they are not concurrent. Therefore, the studies represent different stages of surgical experience. In addition, the patient mix differs significantly, as shown in Table 8. Our series demonstrates a trend toward more operations for female patients, older patients, and nonelective procedures. The populations of patients, indications for operation, and improvements in care are changing the dimension Table 6. Comparison of Mortality after lsolated Primary CABG by Age Reference < 6 Years Old 2 6 Years Old CASS [2] 1.7% (19/6,38) 3.9% (1412,683) 1974-1979 Cleveland Clinic [22] 198-1982 Sutter [this study].5% (18/3,927).6% (21333) 1.2% (3713,173) p <.1 2.1% (12158) p <.1 1984-1985 The p values represent comparison with CASS CABG = coronary artery bypass grafting; CASS = Coronary Artery Surgery Study.

63 Junod, Harlan, Payne, et al: Preoperative Risk Assessment in Cardiac Surgery Table 7. Comparison of Mortality after lsolated Primary CABG by Sex Statistical Reference Male Female Significance" CASS [2] 1.97% (15/7,624) 4.61% (63/1,367) p <.1 1974-1979 Stanford [6].6% (2/362) 1.3% (1/76) NS 1977-1979 Cleveland Clinic [22].6% (35/6,37) 1.9% (211,68) p <.1 198-1982 Sutter [this study] 1984-1985 1.3% (9/681) 2.2% (5/232) NS "The p values are for the male versus female populations CABG = coronary artery bypass grafting; CASS = Coronary Artery Surgery Study; NS = not significant to p <.5 Table 8. Comparison of Patient Populations Having lsolated Primary CABG" % Elective % Emergency Reference % Female % > 7 Years Old Priority Priority CASS [2] 15.2 1974-1979 Stanford [6] 17.4 1977-1979 Cleveland Clinic [22] 15. 198-1982 Sutter [this study] 1984-1985 ~~ 3.4 NA 9.1 25.4 ( p <.1) 25.4 ( p <.1) 79.5 NA 96.8 3.2 98.8 1.2 36. ( p <.1) 14.7 ( p <.1) "The p values represent comparison of Sutter series with others. CABG = coronary artery bypass grafting; CASS = Coronary Artery Surgery Study; NA = not applicable. of cardiac surgery. To give patients a better understanding of risks requires continued examination of the results of operative treatment. A preoperative assessment program helps to quantify the factors. By using the predicted operative mortality as a guide, quality-assurance requirements can be met and performance can be assayed for improvement. Preoperative risk assessment is, therefore, an effective method for quality assurance. References 1. Kennedy JW, Kaiser GC, Fisher LD, et al: Multivariate discriminant analysis of the clinical and angiographic predictors of operative mortality from the Collaborative Study in Coronary Artery Surgery (CASS). J Thorac Cardiovasc Surg 8:876, 198 2. Kennedy JW, Kaiser GC, Fisher LD, et al: Clinical and angiographic predictors of operative mortality for the Collaborative Study in Coronary Artery Surgery (CASS). Circulation 63:793, 1981 3. Fisher LD, Kennedy JW, Davis KB, et al: Association of sex, physical size and operative mortality after coronary artery bypass in Coronary Artery Surgery Study (CASS). J Thorac Cardiovasc Surg 84:334, 1982 4. Hochberg MS, Levine FH, Daggett WM, et al: Isolated coronary artery bypass grafting in patients seventy years of age and older: early and late results. J Thorac Cardiovasc Surg 84:219, 1982 5. Far RS, Golden MD, Javid H, et al: Coronary revascularization in septuagenarians. J Thorac Cardiovasc Surg 86: 616, 1983 6. Miller DC, Stinson EB, Oyer PE, et al: Discriminant analysis of the changing risks of coronary artery operations: 1971-1979. J Thorac Cardiovasc Surg 85:197, 1983 7. Hochberg MS, Parsonnet V, Gielchinsky I, Hussain SM: Coronary artery bypass grafting in patients with ejection fractions below forty percent. J Thorac Cardiovasc Surg 86:519, 1983 8. Loop FD, Lytle BW, Gill CC, et al: Trends in selection and results of coronary artery reoperations. Ann Thorac Surg 36:38, 1983 9. Cobanoglu A, Freimanis I, Grunkemeier G, et al: Enhanced late survival following coronary artery bypass graft operation for unstable versus chronic angina. Ann Thorac Surg 3752, 1984 1. Miller DC, Stinson EB, Rossiter SJ, et al: Impact of simultaneous myocardial revascularization on operative risk, functional result, and survival following mitral valve replacement. Surgery 84:848, 1978 11. DiSesa VJ, Cohn LH, Collins JJ Jr, et al: Determinants of operative survival following combined mitral valve replacement and coronary revascularization. Ann Thorac Surg 34:482, 1982 12. Nunley DL, Grunkemeier GL, Stam A: Aortic valve replacement with coronary bypass grafting. J Thorac Cardiovasc Surg 85:75, 1983

64 The Annals of Thoracic Surgery Vol 43 No 1 January 1987 13. Lytle BW, Cosgrove DM, Loop FD, et al: Replacement of aortic valve combined with myocardial revascularization: determinants of early and late risk for 5 patients, 1967-1981. Circulation 68:1149, 1983 14. Geha AS, Francis CK, Hammond GL, et al: Combined valve replacement and myocardial revascularization. J Vasc Surg 127, 1984 15. Cohn LH, Allred EN, DiSesa VJ, et al: Early and late risk of aortic valve replacement. J Thorac Cardiovasc Surg 88:695, 1984 16. Lytle BW, Cosgrove DM, Gill CC, et al: Mitral valve replacement combined with myocardial revascularization: early and late results for 3 patients, 197 to 1983. Circulation 71:1179, 1985 17. Scott WC, Miller DC, Haverich A, et al: Operative risk of mitral valve replacement: discriminant analysis of 1329 procedures. Circulation 72:Suppl 2:18, 1985 18. Magovern JA, Pennock JL, Campbell DB, et al: Risks of mitral valve replacement and mitral valve replacement with coronary artery bypass. Ann Thorac Surg 39:346, 1985 19. Scott WC, Miller DC, Haverich A, et al: Determinants of operative mortality for patients undergoing aortic valve replacement. J Thorac Cardiovasc Surg 89:4, 1985 2. Myers WO, Davis K, Foster ED, et al: Surgical survival in the Coronary Artery Surgery Study (CASS) registry. Ann Thorac Surg 4:245, 1985 21. Passamani E, Davis KB, Gillespie MS, et al: A randomized trial of coronary artery bypass surgery: survival of patients with a low ejection fraction. N Engl J Med 3121665, 1985 22. Cosgrove DM, Loop FD, Lytle BW, et al: Primary myocardial revascularization: trends in surgical mortality. J Thorac Cardiovasc Surg 88:673, 1984 23. Montague NT 111, Kouchoukos NT, Wilson TAS, et al: Morbidity and mortality of coronary bypass grafting in patients 7 years of age and older. Ann Thorac Surg 39:552, 1985 Discussion DR. BRUCE w. LYTLE (Cleveland, OH): I am grateful for the opportunity to review the manuscript and to discuss this valuable and interesting study. In regard to the specifics of the study, I would ask Dr. Junod to expand a bit on how he and his colleagues used the data from the recent literature to calculate an assigned risk for their patient subgroups. Many studies differ as to the variables implicated in determining mortality and in methods of analysis, whether univariate or multivariate, and I am curious as to how the authors managed to combine data from different studies into a formula for assigning risk. This type of investigation is important. It helps us to focus on high-risk subsets, evaluate the results, and, it is hoped, improve our treatment. But in addition to our interest and the interest of our patients, there is pressure from other sources, including the government, the insurance industry, and the world of communication, to produce figures that may be used in the prediction of risks and in making judgments regarding quality of care. It is important for us to keep in mind and for us to help them to keep in mind that assignment of risk on the basis of historical data is by its very nature imprecise, even when done as carefully as Dr. Junod and his colleagues have done. First, within an institution, indicators of risk will change with time because of changes both in the operative population and in surgical technique and experience. Second, indicators of risk will differ from institution to institution, although one of the strengths of this study is that it does document that major variables have some predictive value even when applied to different institutions. Third, statistics apply to groups of individuals, not to single individuals, and whereas risk can be predicted with some degree of accuracy for a group of patients, for the individual, operative risk is either zero or loo%, and, even for very low-risk subgroups, that risk is not absolutely zero. Fourth, the accuracy of prediction depends in large part on sample size. Despite the excellent results shown here for older patients, I have a hard time accepting the proposition that a wide spectrum of patients with coronary artery disease who are more than 7 years of age can undergo operation with a risk equivalent to that of patients who are less than 7 years old. Whether the differences are apparent and statistically significant is likely to depend on sample size. My last point has to do with the entire concept of quality assurance. There is a tendency, particularly among individuals who are not thoracic surgeons, to equate quality with operative risk. We all know that is just not the case, particularly in regard to elective CABG where risk for most subsets is extremely small. The major issue is the long-term result. This is a factor that must be considered in any calculation of the quality of any treatment of coronary artery disease, and we must continue to impress this point on those who have an observational interest in our specialty. DR. JUNOO: I thank Dr. Lytle and his colleagues at the Cleveland Clinic for their fine work in determining risk and for the reports they have produced. Dr. Lytle asked about our methods of risk determination. We have used recent studies, among which were a number from the Cleveland Clinic. The basis for different assignment is stated in the report and supporting publications are cited. We certainly agree that risks change over time..we have altered some of our risk assignments as a result of this study. Since this study, we have eliminated age and sex as risk factors in isolated primary CABG except for women more than 7 years old. It is likely that the basis for our system of risk assignment will continue to evolve.