Late survival after abdominal aortic repair: Influence of coronary artery

Transcription

1 Late survival after abdominal aortic repair: Influence of coronary artery aneurysm disease L. H. HoNer, M.D., G. Plate, M.D., P. C. O'Brien, Ph.D., F. J. Kazmier, M.D., P. Gloviczki, M.D., P. C. Pairolero, M.D., and K. J. Cherry, M.D., Rochester, Minn. To evaluate long-term survival in relation to preoperative risk factors, we reviewed 1112 patients undergoing abdominal aortic aneurysm (AAA) repair from 1970 to A 6-to 12-year follow-up was obtained on 1087 patients (97.7%) by chart review, death certificates, autopsy reports, and questionnaires returned by patients and referring physicians. Preoperatively 24% of patients had a history of prior myocardial infarction, 19.9% had a history of angina, and 40.4% were hypertensive. Emergency operation for ruptured aneurysm was performed in 6.5% and for expanding aneurysm in 3.4% of patients. The survival rate at 5 years was 67.5% and at 10 years was 40.7%. Cardiac-related problems were the most frequent cause of death (38%); 23% died of myocardial infarction and 15% from other heart disease or sudden death. Other causes included neoplasm (14.6%), other ruptured aneurysm (8.2%), and stroke (6.8%). Cause of death was unknown in 19.6%. A significant correlation of reduced survival time was noted in patients with advanced age and those with evidence of heart disease or hypertension. For patients without preoperative evidence of heart disease or hypertension, the 5-year mortality rate from myocardial infarction was 3.7%, compared with 11.7% for those with a positive history of hypertension and heart disease (p = ). For patients with no preoperative evidence of hypertension or heart disease, the length of survival after AAA repair was the same as that expected for the general population with the same age and sex composition. This study supports the contention that coronary angiography and prophylactic coronary bypass grafting should be performed selectively. Decisions regarding the need for coronary revascularization should be based on symptoms, noninvasive testing, and selective coronary angiography because aneurysmal disease alone is not shown in this study to increase the risk of death from myocardial disease. For patients with clinical findings of coronary artery disease, an aggressive diagnostic approach appears to be justified. (J VAse SURG 1984; 1:290-9.) Abdominal aortic aneurysm (AAA), unlike other atherosclerotic-related diseases, has shown a steady increase over the past 30 years and currently affects over 6% of the U.S. population over the age of65. I In 1950 Estes 2 showed that patients with an infrarenal AAA had a markedly shortened life expectancy compared with the population at large. DeBakey et al? in 1964 and Szilagyi et al. 4 in 1966 clearly demonstrated that patients undergoing aneurysm repair survived longer than patients whose aneurysms were managed nonoperatively. However, it was also noted that the survival time of these surgical patients was still less than that of the general From the Department of Surgery, the Department of Medical Statistics and Epidemiology, and the Department of Internal Medicine, Mayo Clinic. Presented at the Thirty-seventh Annual Meeting of the Society for Vascular Surgery, San Francisco, Calif., June 17-18, Reprint requests: L. H. Hollier, M.D., Department of Surgery, Mayo Clinic, Rochester, MN population. It has been subsequently shown by many authors ~ 14 that coronary artery disease was the major cause of death in these patients. Stimulated by findings of a high incidence coronary occlusive lesions in patients with AAA, 1~ surgeons in several centers performed prophylactic coronary artery revascularization in the hope of decreasing operative mortality rates and improving length of survival. 1 ''2,~G Patients undergoing AAA repair after successful coronary artery bypass grafting clearly have a lower operative mortality rate than those without coronary artery bypass grafting in some institutions, 12 but these patients represented a cohort who often had less severe coronary artery disease and who had survived coronary artery bypass surgery. Approximately 10% of those patients with AAA were believed to have nonreconstructible coronary artery disease and received neither coronary artery revascularization nor AAA repair. 12 In 1980 Brown et al? showed that elective AAA repair 290

2 Volume 1 Number 2 March 1984 Abdominal aortic aneurysm repair 291 Table I. Cause of late death after successful AAA repair Cause No. % of deaths Myocardial infarction Cardiovascular collapse Stroke Other ruptured aneurysm Neoplasia Pulmonary embolism Pneumonia Other known causes Unknown could result in a low mortality rate without routine coronary angiography or prophylactic coronary artery bypass grafting. Routine coronary revascularization appeared to be unjustified except in patients "~:ith symptomatic coronary artery disease. It remained our belief that noninvasive clinical criteria were adequate to detect those patients at greatest risk of myocardial infarction during AAA repair; our operative mortality rate from myocardial infarction for these patients without a history of coronary artery disease was 0.8%. ~ Hertzer, 17 in a 1983 review of over 1000 vascular patients undergoing routine coronary angiography, similarly concluded that most patients with significant coronary artery disease could be idcntificd by conventional clinical criteria. However, there remained the question of longterm survival. Does routinc coronary artery bypass grafting prolong life expectancy in patients with ancurysm? The answer to this question is not readily available, but the following points should be noted: (1) as mentioned previously, some patients with an- ~rysm have such diffuse coronary disease and poor ventricular function that myocardial revascularization is not feasible; (2) the mortality rate from atherosclerotic coronary artery disease in thc United States has been decreasing ovcr the past 10 years independent of any surgical influence18; and (3) although coronary artery bypass grafting appears to have improved the survival time of patients with left-main and three-vessel coronary artery disease, the CASS study, '9 thc European coronary artcry study, 2 and the Duke study 2~ have not yet shown an improved survival time for patients with less extensive asymptomatic coronary artery disease. In view of this information, recent rcports of improved survival time after prophylactic coronary artery bypass grafting and AAA repair cannot be compared with nonconcomitant reports on survival time after AAA Table II. Differences among long-term survival curves of several risk factor categories (death from all causes) Risk factor Age <0.001 Heart disease <0.001 Hypertension Angina pectoris Previous myocardial infarction Surgical indication Year of surgery 0.3 Coronary bypass 0.3 Renal revascularization 0.4 Tobacco use 0.4 repair alone. Two studies done at different times may not be truly comparable, and hence reasonable validity may not be guaranteed. We all agree that significant symptomatic coronary artery disease requires further evaluation and possibly coronary artery bypass grafting prior to elective AAA repair. However, would we improve the survival time of patients with AAA if they all underwent routine prophylactic coronary artery bypass grafting? In other words, in a patient with no clinical history of coronary artery disease, does the mere presence of an AAA indicate a greater risk of death from heart disease than exists in the population at large? In addition, can one identify a subset ofaneurysm patients who are at greater risk of death from myocardial infarction and in whom coronary angiography and myocardial revascularization could be expected to prolong life? METHOD OF STUDY In an attempt to answer some of these questions, we performed a retrospective analysis of clinical findings, operative mortality rates, and length of survival of all 1112 patients undergoing AAA repair at Mayo Clinic over the 6-year period from January 1, 1970, to December 31, Information was obtained by chart review of all patients and by questionnaires and/or telephone interviews with the patient or patient's family and with the referring physician. Death certificates were obtained for patients who died, and autopsy reports were obtained for all patients who underwent postmortem examination. Complete follow-up of 6 to 12 years was available for 1087 patients, or 97.7%. Variablcs reviewed included sex, year of surgery, indication for surgery (elective, urgent, or rupture), preoperative risk factors, operative death, long-term survival, and cause p

3 292 Hollier et al. Journal of VASCULAR SURGERY Table III. Differences among long-term survival curves of several risk factor categories death from all causes, death from MI or CV collapse, and death from MI only) Risk factor All causes MI / CV collapse MI only Age < Hypertension >0.20 Angina pectoris <0.001 <0.001 Previous MI <0.001 <0.001 Angina or prior MI <0.001 <0.001 Other cardiac abnormality < <0.001 (CHF, arrhythmia) Any heart disease <0.001 <0.001 <0.001 MI myocardial infarction; CV = cardiovascular; CHF = congestive heart failure. of death. Particular attention was given to preoperative findings of hypertension, angina, and prior myocardial infarction. Excluded were all patients who had had previous surgery for AAA. Detailed long-term survival curves and frequency distributions for multiple variables were computer generated. Table I provides the distribution of causes of late death for the 621 patients (58.3%) known to have died up to 2 years after successful aneurysm repair. Separate analyses are provided for death from all causes, death from myocardial infarction only (with other deaths considered as censored observation), and death from any cardiac-related cause. Deaths from these variables are analyzed additionally in relation to isolated and multiple preoperative risk factors, such as the year of surgery, tobacco use, concomitant renal revascularization, age, angina, prior myocardial infarction, and hypertension (Table II). The relationship between specific risk factors and survival time is investigated in several manners. Risk factors are individually and jointly analyzed as they relate to overall survival time and to death from all causes, death from myocardial infarction only, and death from any cardiac-related event (Table III). Survival curves were obtained by the Kaplan-Meier method 22 for each subgroup of patients, with comparison among subgroups based on likelihood ratio tests with the Cox model. 8 Association between continuous variables (such as age) and survival time was based on either the Cox regression model or a logic-rank procedure. 23 Comparisons with expected survival time were based on Hyde's procedure. 21 PATIENT INFORMATION Computer analysis provided frequency distributions for each of the variables in the study. Of the patients undergoing aneurysm repair, 974 (87.6%) were men. Age at operation ranged from 31 to 87 years in men and from 27 to 91 years in women (overall median age 68 years). ;~ Tobacco use was prevalent in 81.3% of patients, and 40.4% were hypertensive. Preoperative cardiac disease was identified clinically in 51.4% of the patients, with 19.9% having angina, 24.0% having a prior history of myocardial infarction, and 26.1% having some other manifestation of cardiac dysfunction, for example, arrhythmia, abnormal ECG, or congestive heart failure. Two hundred thirty patients (20.7%) had hypertension and no cfinical evidence of coronary disease, and 342 patients (30,7%) had coronary artery disease with no hypertension. Two hundred seventeen patients (19.5%) had both hypertension and coronary artery disease. Two hundred ninety-eight patients (26.8%) had neither hypertension nor heart disease at the time of aneurysm repair. The operations performed in each of the 6 years were roughly equal in number. Thirty-four patien,~ (3.1%) had concomitant renal artery surgery, and eight of these had bilateral renal artery revascularization. Coronary artery bypass had been performed on only 17 patients (1.5%) prior to AAA repair. Only 20 patients (1.8%) had coronary artery bypass grafting after AAA repair. All coronary artery bypasses were done for severely symptomatic coronary artery disease. Elective aneurysm resections were performed for 998 patients (90.1%), whereas 38 patients (3.4%) had surgery after a diagnosis of acutely expanding aneurysm and 72 patients (6.5%) after rupture of the aneurysm. GENERAL SURVIVAL TIME One thousand sixty-six patients survived surgical repair of the AAA, an overall 30-day operative mortality rate of 4.1%. Surgery for ruptured AAA car-

4 Volume 1 Number 2 March 1984 Abdominal aortic aneurysm repair 293 loo (96.11 ~.~. Over All 6o. ~''''~-~ (79.6) (90.9) ~ "''''~.~ (58.01 ~ 60 i (67.5) ~.... 4O Expected 20 ~ Observed (n : P < o (40.7) Years Fig. 1. Survival curves depicting a comparison between observed survival time, from AAA repair to death from all causes, and expected survival time of the age- and sexmatched general population. loo 8o ~ 60 } 'e. ~ 4o 2O Expected "',.. p ~ Nonruptured {n=1009) "... ~.... Ruptured (n=53) "'-.. I 2 I, I 6 I s,o ' 1'2 I I I Years Fig. 2. Survival time of patients undergoing elective aneurysm repair compared with that of patients undergoing repair of ruptured aneurysms. Operative death is excluded; death from all other causes is included. ried a 26.4% mortality rate, whereas the mortality race for both urgent and elective resection was 2.6% in each case. When deaths from all causes are considered, the probability of surviving 1, 5, and 10 years after successftd surgery is 90.9%, 67.5%, and 40.7%, respectively. These statistics assume that survival distribution in patients lost to follow-up or in those who have had a more recent operation will be the same as in those whose information is available. Median survival time is about 7.25 years. In addition to describing survival time, an expected survival curve was completed to compare survival time preceding all causes of death in this group with survival time in an age- and sex-matched cohort based on west-north central United States life tables. Survival at 5 years is about 84.8% of that expected, and at 10 years observed survival is about 70.2% of that expected (Fig. 1). Available expected ~vival curves are not appropriate for comparison with the disease-specific survival rates. RELATIONSHIPS OF RISK FACTORS TO SURVIVAL TIME Among the risk factors investigated, all but sex, year of surgery, use of tobacco, and renal revascularization appear to influence survival time (Table *II). All the variables measuring preoperative cardiovascular problems were found to decrease survival time significantly. Even excluding hospital deaths, significantly longer survival time was found among patients undergoing elective aneurysm repair than among those undergoing urgent repair or repair of ruptured aneurysms (p = 0.046) (Fig. 2). As expected, survival for all patients is strongly related to the age of the patient (p < 0.001) (Fig. 3, A and B). When deaths from myocardial infarction only are examined, however, there is no statistically significant association with age (p > 0.1). Within each age group, patients with heart disease had shorter survival times than did those without heart disease. This difference was significant in both groups when deaths from all causes and those from myocardial infarction were analyzed. Preoperative hypertension is associated with decreased survival time (p < 0.001) when all deaths are analyzed, but preoperative hypertension appears to have no influence on death from myocardial infarction only (p > 0.1). The overall survival time appears slightly better for those patients with hypertension only than for those with heart problems only, but the difference is not statistically significant (Fig. 4). Patients with a preoperative myocardial infarction, preoperative angina, or both appear to have a shorter survival time than those without either condition (none to either p = 0.012; none to both p = 0.005) (Fig. 5). The amount of time between a preoperative myocardial infarction and operation has been noted to affect hospital mortality rates 24 but did not affect long-term survival (p > 0.1). Seventeen patients had preoperative coronary artery bypass grafting. No statistically significant association was observed between prior coronary bypass and long-term survival (p > 0.2), but the reason may be the small sample size. Patients with both preoperative heart disease and hypertension have significantly shorter survival times than those with either condition alone (p < 0.005) or with neither condition (p < 0.001) (Fig. 6, A). For patients without preoperative evidence of heart disease or hypertension, the 5-year mortality rate from documented myocardial infarction was 3.7%, compared with 11.7% for those with a positive his-

5 294 Hollier et al. Journal of VASCULAR SURGERY 100~.~..,7) (92 8) Age < 60 Years Exp~ted p 0,001 (~) o,oo ~.(97.1) " ~'''-... "''''-.. (84.3) Age 6g-6g Years ~ Age > 70 Years. (72.5) ~.~"- ao """ -'''"--.. (s5.5),oo,(o3.g)... -~ (6g2) ~ ~40 ~ (4s.5) ~ ~,o ~ Ob.rv*d(n:471) ~n-420) ~);pecte~ (420)... p<o.o0~ (28.5) (2g.5)~'~ Years Years Fig. 3. Three survival curves demonstrating influence of age on long-term survival after AAA repair compared with survival of the age- and sex-matched general population (death from all causes was included). Note relative lessening of survival difference in patients over the age of 70 years ~ so ~,o Expected - - Hypertension alone (n=216)... Heart disease alone (n=334) I I I I [ '2 Years Fig. 4. Influence of hypertension alone and heart disease alone on long-term survival, from successful aneurysm repair to death from all causes. tory (p = 0.001). This same result was observed for all cardiac-related deaths (p < 0.001) with 5-year survival rates of 77.3% and 66.8%, respectively. Perhaps most importantly, however, in those patients who had no hypertension and no clinical evidence of heart disease prior to AAA repair, the overall long-term survival remained essentially no different (p > 0.4) from that of the age-matched population at large (Fig. 6, B). DISCUSSION Retrospective analysis of data is fraught with many potential deficiencies. Of particular concern in survival studies is the reliability of determining factual cause of death. For verification of medical information in this study, autopsy reports and death certificates were obtained and clarification of any ambiguity was sought from the attending physician and family. Although errors in diagnosis of death may occur, they have been minimized as much as possible. The age and sex distribution is comparable to other, similar reports in the literature, 1)'13"14 an~ these patients reflected a similar incidence and d~tribution of risk factors. As in other studies,* coronary artery disease continues to be a major factor in postoperative death and long-term survival, but subgroups with greater and lesser risks can be developed. In addition, noncardiac risk factors have varying importance in long-term survival. Somewhat surprisingly, smoking did not appear to adversely influence late mortality rates in this series. A similar lack of adverse effect of smoking on the incidence of angina pectoris and the risk of myocardial infarction has also been reported from the Framingham study for persons over 50 years of age. z~ As one might expect, actual survival time after aneurysm repair is very strongly related to the age of, the patient. However, if one evaluates length of *References 3-5, 11-14, 16.

6 Volume 1 Number 2 March 1984 Abdominal aor-tic aneurysm repair 295 Previous Myocardial Infarction Hypertension and Heart Disease Expected -0bser ed (n=20,, (20.5) Angina PeCtOriS Expected -NO angina, =*a,.. ( S 207, P = ; I a L ; k Ib L, * Year5 Fig. 5. Influence of previous myocardial infarction, A, and angina pectoris, B, on long-term survival, from successful AAA repair to death from all causes. survival in relation to expected survival time for specific age groups, the difference appears to diminish with advancing age (69% of expected for patients less than 60 years of age compared with 72% of expected for those older than 70 years of age), although it is not statistically significant. Of particular interest is the finding that 11.8% of patients who survived AAA repair subsequently died of rupture of another aneurysm or a late complicaan of the original aneurysm repair. Crawford and Cohen9 previously noted the prevalence of additional aneurysms in other segments of the aorta and the high incidence of late development of other aneurysms. These factors suggest that patients who undergo aneurysm repair should be followed up closely for the development of other aneurysms over the ensuing years. Although B-mode ultrasound scanning is a useful screening tool for AAAs, it is somewhat less effective in delineating thoracic, suprarenal, and iliac aneurysms unless special studies are performed. We have found computerized tomography to be much more definitive for these particular areas and would suggest that it be used for complete aortic evaluation. Hypertension significantly reduced survival time but had no influence on death from myocardial infarction alone. Death from stroke, however, was Fig. 6. A, Influence of combined hypertension and heart disease on long-term survival, from successful AAA repair to death from all causes. B, Long-term survival of patients who have no hypertension and no overt heart disease following successful AAA repair. This is the same as the expected survival time of an age- and sex-matched general population. more commonly associated with patients having hypertension, as was death from rupture of other aneurysms. In addition, hypertension further decreased survival time in those patients who also had clinical evidence of coronary artery disease (p < 0.005). DeBakey et a1.3 in 1964 similarly noted the differential risks in patients with hypertension and heart disease. It appears, therefore, that strict control of hypertension could potentially improve long-term survival rates by decreasing mortality rates associated with myocardial infarction, stroke, and aneurysmal rupture. Obviously, there is a need for further clarification of the interrelationship of hypertension with other risk factors in patients with AAA. Only 17 of the 1112 patients in this study had undergone prior coronary bypass. After repair of the aneurysm, at least 181 patients (16.3%) had a nonfatal myocardial infarction sometime after surgery, and at least 254 patients (22.8%) had angina in the late postoperative period. Nevertheless, only 20 of these patients having aneurysm repair between 1970 and 1976 underwent subsequent myocardial revas-

7 296 Hollier et al. Journal of VASCULAR SURGERY cularization. Although one would expect that patients who had coronary revascularization would have a prolonged survival time, such was not the case. There may be several explanations. One possible explanation is that other previous reports showing increased survival time after myocardial revascularization used nonconcurrent controls for comparison. ~,'2"~,~ Because there has been a general reduction in cardiovascular-related deaths in this country over the past 10 to 12 years, ~s the improved survival time of coronary bypass patients may merely reflect recent general trends in the population. Because our study reports time-matched controls, this false improvement in survival time would not be present. The most likely explanation, however, is the very small number of patients in our study who underwent coronary artery bypass grafting. It appears likely that we had an insufficient data base for valid comparison and thus could not demonstrate the improvement in long-term survival in the coronary bypass patient. Both the CASS study TM and the European coronary artery study 2 clearly demonstrate the superiority of coronary bypass over medical treatment for symptomatic coronary artery stenosis. AAAs continue to represent significant morbidity and mortality rates in our elderly population. Although aneurysm repair clearly prolongs life, the long-term survival of these patients following successful surgery remains less than the survival time of the population at large, with the 10-year survival rate being only 70% of the expected rate. Our study, as well as others, '~'~-~4 has documented coronary artery disease as the major cause of death in these patients. It has been suggested that coronary revascularization may improve both the operative and long-term survival of these patients. We have previously demonstrated that aneurysm repair can be accomplished safely without prophylactic coronary revascularization 5 and that delaying aneurysm repair appears to result in an increase in aneurysm rupture. Crawford et al. ~ have noted that "a significant number of patients ruptured their aneurysm while recovering from their bypass operation"; these authors further suggested performing prophylactic coronary bypass surgery only in patients with impending infarction and treating the other patients medically until a later date, after recovery from aneurysm operation. ~' We have also used this approach. However, even after patients have recovered from aneurysm repair, we have generally reserved coronary revascularization for patients with limiting symptoms. In view of the markedly reduced survival time of patients with symptoms of coronary artcry disease, however, a more aggressive approach may be warranted. Despite the attractiveness of that line of reasoning, prophylactic coronary revascularization, even if limited only to patients with clinical evidence of coronary artery disease, has somc limitations. Because coronary artery bypass grafts do have a significant incidcnce of graft occlusion, v and reoperative coronary revascularization carries a somewhat higher mortality rate and reduced survival time, 26 routine coronary bypass of early asymptomatic lesions may put the patient at greater risk of myocardial infarction in the long run should graft failure occur. At present it appears that coronary artery bypass grafting for only mildly symptomatic onc- or two-vessel coronary artery disease (except left-mas.~ coronary artery disease) does not prolong life 6"~J~ and should not be undertaken for asymptomatic patients. A more effective approach might be to attempt to more clearly delineate those patients at greatest risk of death from myocardial infarction. Clearly, patients with no hypertension and no clinical evidence of heart disease have a normal life expectancy and are least likely to benefit from prophylactic coronary angiography and coronary bypass. On the other hand, patients with severe angina generally derive great benefit from myocardial revascularization. Those patients who have clinical evidence of stable coronary artery disease are more likely to demonstrate coronary stenosis on angiography, but not all would benefit from coronary bypass. 17 Further evaluation, however, is clearly indicated. Coronary angiography can identify and quantitare coronary artery stenosis, but the presence ofsteno}fs does not necessarily imply that myocardial infarction is likely, especially if extensive collaterals have developed. Stress ECGs seem to be insufficiently sensitive to identify all angiographic lesions, although the importance of those lesions might be questioned if they do not produce signs of ischemia with exercise. Multigated radionuclide scans appear to be useful in screening patients with coronary artery disease because they can provide quantitation of ventricular function and can identify wall motion abnormalities occurring with exercise. Left ventricular ejection fraction appears to correlate well with both operative mortality rates and long-term survival 6 and appears to have a high correlation to coronary lesions identified angiographically that pose significant risk of myocardial infarction. In fact, left ventricular function seems to be a more important

8 Volume 1 Number 2 March 1984 Abdominal aortic aneurysm repair 297 prediction of survival time than does the number of diseased vessels. ~ Perhaps the best method of evaluating stable patients for high risk of myocardial death is a combination of history and physical examination, ECG, multigated radionuclide scan of patients with findings suggestive of coronary artery disease, and coronary angiography limited to panents who have evidence of functional myocardial ischemia. An aggressive diagnostic approach certainly appears to be justified for patients who have clinical evidence of coronary artery disease. However, myocardial revascularization should perhaps be reserved for those patients with significant symptoms or life-threatening coronary lesions. REFERENCES "% BickerstaffLK. Hollier LH, Van Peenen HJ, Melton III LJ, Palrolero PC, Cherry KJ. Abdominal aortic aneurysms: The changing natural history. J VASC SURG 1984; 1: Estes Jr JE. Abdominal aortic aneurysm: A study of one hundred and two cases. Circulation 1950: 2: DeBakey ME, Crawford ES, Cooley DA. et al. Aneurysm of abdominal aorta: Analyses of results of graft replacement therapy one to eleven years after operation. Ann Surg 1964: 160: Szilagyi DE. Smith RE. DeRusso FJ, et al. Contribution of abdominal aort,c aneurysmectomy to prolongation of life. Ann Surg 1966: 164: Brown OW. Hotlier LH, Palrolero PC, et al. Abdominal aortic aneurysm and coronary artery disease: A reassessment. Arch Surg 1981: 116: Califf RM, Tomabechi Y, Lee KL, et al. Outcome in onevessel coronary artery disease. Circulation 1983; 67: Campeau L. Lesp4rance J, Hermann J, et al. Loss of the improvement of angina between one and seven years after aortocoronary bypass surgery. Circulation 1979: 60:I1-I5. 8. Cox DR. Regression models and life tables. JR Statistical Soc 1972: 34: *,... 9 Crawford ES, Cohen ES. Aortic aneurysm: A multi-focal disease. Arch Surg i982; 117: Crawford ES. Morris Jr GC, Howell JF, et al. Operative risk in patients with prev,ous coronary artery bypass. Ann Thorac Surg 1978; 26: Crawford ES, SalWa AS, Bagg III JW, et al. Infrarenal abdominal aortic aneurysm: Factors influencing survival after operation performed over a 25-year period, Ann Surg 1981; 193: , Hertzer NR. Fatal myocardial infarction following abdominal aortic aneurysm resection, Ann Surg 1980; 192: , Johnson Jr G, Gurri JA, Burnham SJ. Life expectancy after abdominal aortic aneurysm repair. In: Bergan JJ, Yao JST, eds. Aneurysms: Diagnosis and treatment. New York: Grune & Stratton, Inc, 1982: Thompson JE, Hollier LH, Patman RD, et al. Surgical management of abdominal aortic aneurysms: Factors influencing mortality and morbidity: A 20-year experience. Ann Surg 1975; 181: Hertzet NR, Young JR, Kramer JR, et al. Routine coronary angiography prior to elective aortic reconstruction. Arch Surg 1979; 114: McCollum CH, Garcia-Rinaldi R, Graham JM, et al. Myocardial revascularization prior to subsequent major surgery in patients with coronary artery disease. Surgery 1977; 81: FIertzer NR. Myocardial ischemia. Surgery 1983; 93: Fcinleib M, Havlik RJ, Thorn TJ. The changing pattern of ischemic heart disease. J Cardiovasc Med 1982; 7: Mock MB, Rengqvist I, Fisher LD, et al. Survival of medically treated patients in the coronary artery surgery (CASS) registry, Circulation 1982; 6& , European Coronary Artery Surgery Group. PrOspective randomized study of coronary artery bypass surgery in stable angina pectoris. Lancet 1980; 2: Hyde J. Testirlg survival under right censoring and left truncation. Biometrika 1977; Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Statistical Assoc 1958; 53: O'Bricn PC. A non-parametric test for association with censored data. Biometrics 1978; 34: Tathan S, Moffitt EA, Taylor WF, et al. Myocardial infarction after general anesthesia. JAMA 1972; 220: Dawber TR. The Framingham study: The epidemiology of atherosclerotic disease. Cambridge, Mass: Harvard University Press, 1980: Loop FD, Cosgrove DM, Kramer JR, et al. Late clinical and arteriographic results in 500 coronary artery operations. J Thorac Cardiovasc Surg 1981; 81: DISCUSSION Dr. George Johnson, Jr. (Chapel Hill, N.C.). Dr. HolLier's data continue to add to the pool of developing knowledge on the longevity of patients on whom vascular surgeons operate. Last year Dr. Steven Burnham reported our late follow-up of patients operated on for AAA and for aortoiliac or femoropopliteal occlusive disease. The longevity following these operations was half that of the general population. The present report suggests that we might find a subset of these patients who are at a high risk for late death and in whom we could alter these risk factors. Stimulated by Dr, Hollier's abstract, we reviewed our experience. Five hundred fifty-six patients have had repair of their aortic aneurysm at North Carolina Memorial Hospital, and 453 of these survived the operation. The longevity of patients with risk factors, including hypertension, diabetes, heart disease, stroke, and azotemia, was compared with that of patients with none of these risk factors. The patients without any risk factors lived longer than those with risk factors, but not as long as the general population. We are surprised that Dr. Hollier was able to identify a group of these patients with manifest arteriosclerosis who lived as long as the control population. \

9 298 Hollier et al. Journal of VASCULAR SURGERY Perhaps we North Carolinians just do not live as long as the rest of you. One thing we did notice in the study (and it is a danger in all studies that look at the cause of late death) is the reliance on the death certificate. Many practicing physicians denote myocardial infarction as the cause of death when the true cause is unknown. For the absolute answer, we will have to await long-term results after coronary artery bypass grafting to sec whether this procedure will prolong survival. Another interesting point that we found in our series, and that Dr. Hollier emphasizes, is the longevity of the aged patient. The older the patient operated on in both series, the nearer the longevity approaches that of the general population. I am sure this is a reflection of patient selection. We agree that a reasonable attempt should be made to identify those patients with major risk factors and to correct the problems if possible before performing operations on patients with AAA. Dr. Norman R. Hertzer (Cleveland, Ohio). In an attempt to reduce the early and late mortality rates from myocardial infarction following peripheral vascular reconstruction, patients under consideration for vascular procedures at the Cleveland Clinic have been advised to undergo preoperative cardiac catherization since As we have emphasized on several occasions, the objective of our study of routine coronary angiography has always been to determine those patients for whom this approach would have the greatest benefit if used on a selective basis. Because our previous experience with survival time after aortic aneurysm resection is so similar to that described from the Mayo Clinic, you might be interested in a couple of results concerning the first 1000 patients included in our current investigation. A total of 302 of the 1000 patients had aortic aneurysms as a primary or incidental diagnosis. On the basis of the cardiac history and a standard ECG, coronary artery disease could be suspected in 158 patients even prior to cardiac catheterization. Severe, surgically correctable lesions were documented by angiography in 42% of this group, and severe, inoperable disease was identified in another 9.5%. No clinical evidence of coronary involvement was present in 144 of the 302 patients. Although surgically correctable coronary disease still was found in 19% and inoperable lesions were found in a few others, these differences were highly significant statistically. Significant differences in the incidence of severe coronary artery disease were also found in other subsets, most notably among men, patients over 60 years of age, and patients with diabetes. Nevertheless, previous symptoms or ECG indications of coronary disease--including ST-T segment changes, by the way--were critical features of correctable disease in all these groups. If patients with aneurysms have clinical evidence of coronary disease, angiography will demonstrate severe lesions in nearly 50%, most of which are amenable to myocardial revascularization. Use noninvasive screening if you will, but restrict it to those in whom coronary disease is not suspected on conventional grounds. Why be concerned about all this? For one thing, the operative mortality rate after aneurysm resection was only 1.6% among 61 patients in our current series who required preliminary myocardial revascularization because of survey coronary angiography. For another, 5-year actuarial survival after aneurysm resection at our center for patients with suspected but uncorrected coronary disease previously has been about 25% worse than for patients of the same age who receive coronary artery bypass. Most of this meaningful difference, of course, is caused by fatal myocardial infarction. We simply believe that many of these deaths can be prevented, and with a little patience, we hope to prove it. Dr. HoNer (closing). In response to Dr. Johnson's question concerning death certificates, we agree that those diagnoses are often in error. However, our study is not based solely on the cause of death, as you noted, but "~ based on death by any cause. Death is the one end poinf about which no statistician can argue. I appreciate your taking the time to again review your data. At this point I find it somewhat difficult to explain your failure to concur with our findings. I might point out that our findings are not entirely new. In i964 Dr. De- Bakey demonstrated that in patients who had no hypertension and no overt heart disease, long-term survival was markedly improved over that of patients who did have clinical hypertension, heart disease, or both. I urge Dr. Johnson to again examine his data and be sure that they are matched with a comparable age- and sex-matched cohort for comparison of the expected survival time. We all are indebted to Dr. Hertzer for the great contributions he has made in delineating the prevalence of coronary artery disease in vascular patients. His data demonstrate an improved operative mortality rate and an improved long-term survival rate in patients who have undergone successful coronary bypass and subsequently) have had aneurysm repair in recent years. However, one must be cautious when comparing current surgical patients with patients who had surgery 15 years ago. Over the past 10 years there has been an overall reduction in death from coronary artery disease in this country, and this reduction appears to be independent of the advent of coronary bypass. In addition, bias in selection occurs against patients who undergo aneurysm repair, since 4% to 10% of aneurysm patients have such severe coronary artery disease that coronary bypass is not undertaken. Although coronary angiography can identify and quantitate coronary stenosis, the presence ofstenosis does not necessarily imply that myocardial infarction is likely. In fact, left ventricular function may perhaps be a more important predictor of survival than is the number of diseased vessels. Further, let us not forget that age is one of the major predictors of the operative mortality rate, particularly for coronary bypass. Multiple reports on the

10 Volume 1 Number 2 March 1984 Abdominal aortic aneurysm repair 299 CASS study clearly demonstrate that the mortality rate for myocardial revascularization is increased after the age of 65 years. In patients over the age of 65 to 70 years, perhaps the issue is not so much prophylactic coronary surgery for survival but the role of this surgery in improving the quality of life for symptomatic patients who have clinical evidence of coronary artery disease and for whom an aggressive diagnostic approach certainly is justified. However, myocardial revascularization of patients in this age group should perhaps be reserved for those with significant symptoms or maly life-threatening lesions.