Early and Late Mortality of Patients Undergoing Aortic Valve Replacement After Previous Coronary Artery Bypass Graft Surgery
Background In a small number of patients who undergo coronary artery bypass graft surgery (CABG), a hemodynamically significant aortic valve lesion requiring aortic valve replacement (AVR) develops as they grow older. In a limited number of studies in small patient groups, high mortality has been shown in patients undergoing AVR after CABG. We undertook this study to determine the mortality risk factors for patients who undergo AVR after CABG procedures.
Methods and Results The outcome of 104 patients treated at our institution between January 1983 and December 1993 was retrospectively reviewed. The initial surgery was CABG in all patients. The patient population included 86 men (83%) and 18 women (17%); their mean age was 67 years. Overall, 70% of patients had congestive heart failure, and 96% had multivessel coronary artery disease. The diagnosis was aortic stenosis in 68% of patients, aortic insufficiency in 16%, and combined aortic stenosis and aortic insufficiency in 16%. Postoperative complications included worsening congestive heart failure (35%), perioperative myocardial infarction (13%), and bleeding (28%). The early mortality was 14%, and the late mortality was 17% (mean follow-up, 35 months). The risk factors for early mortality were number of diseased vessels (P=.028), renal failure (0.000), and prior myocardial infarction (P=.028). A perioperative predictor of early mortality was cardiopulmonary bypass time (P=.000). The risk factors for late mortality included preoperative diabetes mellitus (P=.007), postoperative acute respiratory distress syndrome (P=.011), and ventricular arrhythmias (P=.0001). The survival at 1, 5, and 10 years was 96%, 75%, and 49%, respectively.
Conclusions Risk factors were identified for early and late mortality in patients undergoing AVR after previous CABG. Although early morbidity and mortality were high, the long-term outcome of the survivors was favorable.
The surgical results and late outcome of patients undergoing isolated aortic valve replacement (AVR) and AVR with concomitant coronary artery bypass graft surgery (CABG) have been well established.1 2 3 4 5 6 7 8 9 10 Little information is available on AVR after CABG, but studies in small groups of patients have shown a very high perioperative mortality associated with this procedure.11 We observed in a small number of patients with prior CABG that a hemodynamically significant aortic valve lesion requiring AVR develops as they grow older. We thus undertook the present study to examine the early and late outcomes of AVR in patients with previous CABG surgery and to identify predictors of early and late outcome in these patients.
The study group comprised 104 patients who had undergone AVR after previous CABG between January 1983 and December 1993 at the Texas Heart Institute. Patients undergoing concomitant mitral valve replacement and those with aortic dissection were excluded from the study. Patients were identified through computer-generated records, and in-hospital outcome was determined by review of the patient charts. Long-term follow-up was obtained through questionnaires and telephone calls.
Clinical and Laboratory Data
Etiologic classification of the aortic valve disease was based on findings at cardiac catheterization, echocardiography, surgery, and pathological examination of the excised valve. On the basis of the clinical examination with echocardiographic and angiographic studies, patients were classified as having pure aortic stenosis (AS), pure aortic insufficiency (AI), or combined AS and AI.
Cardiac catheterization and coronary angiography were performed routinely before all cardiac operations. Narrowing of the left main coronary artery by more than 50% and narrowing of the other coronary artery branches or bypass grafts by more than 70% were considered significant.
Perioperative myocardial infarction was defined either by the appearance of a new Q wave on the postoperative ECG or when the postoperative MB isoenzyme of creatine kinase was more than 50 units and comprised more than 10% of the total creatine kinase in patients with equivocal ECG changes.
Renal failure was defined as serum creatinine of more than 2.0 mg/dL. Left ventricular ejection fraction was derived from one or more procedures: left ventricular angiography, two-dimensional echocardiography, and left ventricular gated blood pool nuclear study.
Early mortality was defined as any death occurring during or within the first 30 days after surgery. An in-hospital death occurring more than 30 days after surgery was also considered early mortality.
Valvular surgery was classified as “urgent” if it had to be performed in patients with clinical or hemodynamic instability. All patients underwent cardiopulmonary bypass with systemic hypothermia and hypothermic blood or crystalloid potassium cardioplegia. Cardioplegic solution was delivered in an antegrade fashion into either the aortic root or the coronary ostia directly with hand-held cannulae when aortic regurgitation was significant. In three patients, the cardioplegic solution was delivered into the coronary sinus. Concomitant CABG was performed with the use of saphenous vein or internal mammary artery bypass grafts. Distal anastomoses were usually performed before valve replacement to permit vein graft perfusion with cardioplegic solution. The selection of the valve prosthesis and the decision to perform CABG were made by the surgeon.
The survivors were compared with the nonsurvivors according to clinical, anatomic, and surgical variables (Table 1⇓). χ2 and Student’s t tests were used for univariate analysis; those with P≤.5 were incorporated into a model of multivariate analysis with the use of sas-stat software. Logistic regression was used for early death with the logistic procedure. Cox regression was used for late death with the phreg procedure. P<.05 was considered significant. A survival curve was constructed with the life test procedure.
The initial surgery was CABG in all patients (Table 2⇓). The mean number of bypasses grafted was 2.9 per patient. The mean time interval between the initial surgery and the index AVR surgery was 9.0±3 years. The mammary artery was rarely used as a bypass graft in the initial surgery.
Clinical and Laboratory Data at Index AVR Surgery
Overall, the patient population was older (mean age, 67 years), and the majority were men (Table 3⇓). The most common symptoms at presentation in order of frequency were congestive heart failure, angina, and hypertension. Prior myocardial infarction was noted in 38 patients. Three patients (3%) had evidence of bacterial endocarditis.
AS was documented in 68% of patients, AI in 16%, and combined AS and AI in 16%. The incidence of one-, two-, and three-vessel coronary artery disease (CAD) was 4%, 15%, and 81%. Left ventricular ejection fraction was severely reduced (≤30%) in 18 patients (22%).
Index AVR Surgery
Surgery was performed on an urgent basis in 3 patients (3%) and as an elective procedure in 101 patients (97%) (Table 4⇓). The operation was AVR in 45 patients (43%) and AVR with concomitant CABG in 59 patients (57%), with a mean of 1.9 bypasses grafted per patient. A mechanical prosthesis was used in the majority of patients (92%). Based on pathological examination, the aortic lesion was determined to be calcific degeneration in 64%, bicuspid aortic valve in 12%, and rheumatic disease in 8% of patients, respectively. Six additional patients (6%) had ectasia of the ascending aorta.
Major postoperative complications occurred in 63 patients (61%) (Table 5⇓). The most common postoperative complications were congestive heart failure, severe bleeding requiring surgical exploration and/or transfusion with more than 4 units of blood, supraventricular and ventricular arrhythmias, low cardiac output, sepsis, perioperative myocardial infarction (MI), acute respiratory distress syndrome, renal failure, and cerebrovascular accident. The diagnosis of perioperative MI was based on ECG criteria in 13 patients and on elevation of the MB isoenzyme of creatine kinase in one additional patient, with evidence of poor R-wave progression in the anterior chest leads. Perioperative MI occurred in two patients undergoing AVR alone and in 12 patients undergoing AVR and concomitant CABG (P=.019).
Early mortality (≤30 days after surgery) was documented in 15 patients (14%), with 3 intraoperative and 12 postoperative deaths. The most common causes of death were congestive heart failure and low cardiac output in 12 patients (80%), ventricular arrhythmias in 3 patients (20%), severe bleeding in 2 patients (13%), cerebrovascular accident in 1 patient (7%), and acute respiratory distress syndrome in 1 patient (7%). Some patients who died early had more than one of these diagnoses.
Predictors of Early Mortality
Univariate analyses were performed based on 36 preoperative and intraoperative variables that were analyzed for possible relation to surgical mortality (Table 1⇑). By univariate analysis, the preoperative predictors of early mortality were unstable angina, prior MI, multivessel CAD, renal failure, and cerebrovascular accident (Table 6⇓). The perioperative univariate predictors were surgery performed on an emergency basis, the need for intra-aortic balloon pump assistance at the end of surgery, and longer cardiopulmonary bypass time.
Preoperative predictors of early postoperative mortality by multivariate analysis were renal failure, prior MI, and multivessel CAD. A perioperative predictor was longer cardiopulmonary bypass time (Table 7⇓).
Follow-up was obtained in 89% of patients. The mean follow-up was 35 months. The late mortality was 17%. The survival rate at 1, 5, and 10 years was 96%, 75%, and 49%, respectively. Early perioperative death was not taken into consideration (Figure⇓).
Multivariate Predictors of Late Mortality
The preoperative predictor of late mortality was diabetes mellitus (Table 8⇓). Postoperative predictors included ventricular arrhythmias and acute respiratory distress syndrome.
Most of the disease processes that commonly involve the aortic valve have been shown to progress in severity with advanced age.12 13 14 These include degenerative calcific aortic valve, rheumatic aortic valve disease, and congenital bicuspid aortic valve. As the cohorts of patients who had CABG over the past 25 years grow older, the incidence of a hemodynamically significant aortic lesion requiring AVR is expected to rise accordingly. Some studies have suggested that concomitant CAD can accelerate the rate of progression of AS.14 The present study was initiated to examine the early and late outcomes of AVR in patients with prior CABG.
The surgical mortality associated with AVR has been progressively decreasing because of improvements in myocardial preservation, surgical techniques, and postoperative care. In most institutions, the risk of early mortality in isolated AVR has been reported to range from 2% to 6%.1 4 5 6 7 8 9 10 The predictors of early mortality include age, left ventricular function, prior aortic valve surgery, aortic insufficiency, and the presence of CAD.1 2 3 4 5 6 7 8 13 Although concomitant CABG may improve the long-term results for patients with aortic valve disease and CAD, it also increases the risk of early mortality.1 5 6 9
AVR and CABG Procedures
Kirklin and colleagues1 2 reported an early mortality of 2.9% in patients with primary isolated AVR, 6.5% with repeat AVR, and 9% with concomitant AVR and CABG. Mullany et al4 compared early mortality in patients with CAD and aortic disease undergoing AVR only with similar patients undergoing AVR and CABG: mortality was 6.3% for AVR alone and 10% for AVR and CABG. Although the difference between these rates was not statistically significant, a definite trend toward higher mortality in the AVR/CABG group was noted.
A group from the University of Toronto7 published their experience with AVR in elderly patients; mortality was 7.6% in patients older than 70 years versus 5.8% in patients younger than 70. The mortality was 11.5% in patients with concomitant CAD treated with CABG and 21% in patients with CAD with no surgical intervention. In the Veterans Administration Cooperative Study, the mortality for AVR was 8.3% and increased to 12.8% in patients with three-vessel CAD.5 The mortality in patients with previous cardiac operations was 30.4%. Collins and Aranki11 reported 44 patients who underwent AVR after previous CABG. The mean interval between both procedures was 68 months, and 24 patients had evidence of mild-to-moderate AS at the time of CABG. The surgical mortality of AVR in this group was 18.2%. Based on this high mortality rate, the authors concluded that concomitant AVR should be performed in patients with evidence of mild-to-moderate AS at the time of CABG.
In the present study, the 104 patients had no evidence of significant aortic valve disease at the time of the initial CABG procedure and the mean interval between both operations (109 months) was longer than that reported in the study of Collins and Aranki.11 This lessens the likelihood that AVR would have been necessary at the time of initial CABG procedures in our patients.
Effect of CAD on Surgical Outcome
It has been demonstrated that CAD decreases the tolerance of the hypertrophied myocardium to ischemia, thereby interfering with myocardial preservation during open heart surgery.1 15 16 In the present study, we had the opportunity to study the effect of patent vein grafts from a prior surgery on myocardial preservation and perioperative MI during AVR. Patent vein grafts from a prior CABG may improve myocardial preservation during AVR by allowing cardioplegic solution to reach distal coronary arterial beds beyond areas of proximal stenosis. Mammary artery grafts, however, do not provide the same protection. In the present study, 14 patients (13%) had evidence of a new transmural MI after AVR. Perioperative MI occurred more frequently when CABG was performed concomitant with AVR. The presence of CAD at the time of AVR procedures has been reported to be associated with a higher incidence of perioperative MI.17 18 19 Iung and colleagues17 reported a perioperative rate of MI of 0.7% after AVR in patients with normal coronary arteries and 6.9% in patients with aortic valve disease and CAD undergoing AVR and CABG. Reed and colleagues19 reported a 7.6% incidence of perioperative MI after AVR and CABG. MI was diagnosed by ECG criteria and/or elevation of the MB isoenzyme of creatine kinase. With ECG criteria, Jones et al18 reported a perioperative MI rate of 17% after AVR and 15% after AVR plus CABG in a group with aortic valve disease and concomitant CAD. Our findings indicate that patent vein grafts from previous CABG do not lower the incidence of perioperative MI at the time of AVR. This study also reveals that patients who do not have concomitant revascularization while undergoing AVR after previous CABG have a lower perioperative MI rate.
Some investigators believe that in the presence of severe CAD, retrograde delivery of cardioplegic solution through the coronary sinus generates better global distribution of cardioplegia and cooling than does antegrade delivery and can help optimize preservation of myocardium in jeopardy during surgery.20 This technique was uncommonly used at our institution at the time of this study.
Predictors of Early Mortality
All five predictors of early mortality by univariate and multivariate analyses in this study are related to variables and conditions already known to affect results of AVR and heart surgery in general. These factors are reported to include age, left ventricular function, associated systemic illnesses, prolonged ischemic time, prior heart surgery, and low cardiac output at the end of the surgical procedure.1 2 3 4 5 6 7 8 9 10 11 Several variables that correlated with early mortality by multivariate analysis, however, were not found to be independent predictors of early mortality by multivariate analysis. These variables were preoperative cerebrovascular accident, preoperative unstable angina, emergency surgery, and the use of IABP after surgery.
Reduced left ventricular contractility has been shown repeatedly to be associated with a poor early and late outcome after AVR, for both AS and AI.1 3 7 8 9 10 The combination of CAD and aortic valve disease can alter ventricular function by different mechanisms, including increased wall stress, myocardial ischemia, and MI.15 16 It is interesting that a prior MI—and not left ventricular ejection fraction—in this study was an independent predictor of poor outcome because areas of MI are not expected to regain function after the aortic lesion is corrected. The finding of increased mortality in the subgroup of patients with multivessel CAD is concordant to reports in which patients with combined aortic valve disease and CAD with a higher number of diseased coronary artery vessels have a higher surgical mortality.5 7 8 An independent perioperative predictor of early mortality was prolonged cardiopulmonary bypass time.
Perioperative MI, however, was not found to be an independent predictor of early mortality.
Other factors that did not significantly affect early mortality included age, nature of the aortic lesion (AS, AI, or combined AS and AI), and concomitant coronary artery bypass revascularization at the time of the index surgery.
The incidence of late mortality in our series was higher than that reported after isolated AVR but similar to that of patients with AVR and concomitant CABG. Kirklin and Barratt-Boyes1 had reported an overall survival of 91% and 77% at 1 and 5 years after isolated AVR. In the series reported by Magovern and colleagues,6 in which 52% of patients had AVR plus CABG, 72% of hospital survivors were alive at 48 months. In a report from the Cleveland Clinic9 of patients with AVR and CABG, the survival rate was 89%, 79%, and 52% at 2, 5, and 10 years, respectively. Previously reported risk factors for late mortality after AVR include old age, poor New York Heart Association functional class, ventricular enlargement or dysfunction, coexistent CAD, mitral regurgitation, and ventricular arrhythmias.1 2 3 4 5 6 7 8 9 10 In our study, the preoperative predictor of late mortality was diabetes mellitus. The postoperative predictors of late mortality were acute respiratory distress syndrome and postoperative ventricular arrhythmias. Age and left ventricular ejection fraction were not factors for late mortality in this study.
In the present study, we identified predictors of early and late mortality in patients undergoing AVR after previous CABG. Early and late outcomes were not affected by the nature and etiology of the aortic valve lesions (AS, AI, or combined AS and AI; degenerative versus other pathological lesions). Although AVR after a previous CABG procedure is associated with a high rate of early morbidity and mortality, the long-term outcome of survivors appears favorable.
- Copyright © 1995 by American Heart Association
Kirklin JW, Barratt-Boyes BG. Cardiac Surgery. 2nd ed. New York, NY: Churchill Livingstone; 1993:524-571.
Kirklin JK, Naftel DC, Blackstone EH, Kirklin JW, Brown RC. Risk factors for mortality after primary combined valvular and coronary artery surgery. Circulation. 1989;79(suppl I):I-185-I-190.
Sethi GK, Miller DC, Souchek J, Oprian C, Henerson WG, ul Hassan Z, Folland E, Khuri S, Scott SM, Burchfiel C, Hammermeister KE. Clinical hemodynamic and angiographic predictors of operative mortality in patients undergoing single valve replacement. J Thorac Cardiovasc Surg.. 1987;93:884-897.
Fremes SE, Goldman BS, Ivanov J, Weisel RD, David TE, Salerno T, and the Cardiovascular Surgeons at the University of Toronto. Valvular surgery in the elderly. Circulation. 1989;80(suppl I):I-77-I-90.
Lytle BW, Cosgrove DM, Goormastic M, Loop FD. Aortic valve replacement and coronary artery bypass grafting for patients with aortic stenosis and coronary artery disease: early and late results. Eur Heart J. 1988;9(suppl E):143-147.
Lund O, Nielsen TT, Magnussen K, Pilegaard K, Knudsen M. Valve replacement for calcified aortic stenosis in septuagenarians infers normal life length. Scand J Thorac Cardiovasc.. 1991;25:37-44.
Collins JJ, Aranki SF. Management of mild aortic stenosis during coronary artery bypass graft surgery. J Cardiac Surg. 1994;9(suppl I):145-147.
Wagner S, Selzer A. Patterns of progression of aortic stenosis: a longitudinal hemodynamic study. Circulation. 1982;65:709-712.
Delaye J, Chevalier P, Delahaye F, Didier B. Valvular aortic stenosis and coronary atherosclerosis: pathophysiology and clinical consequences. Eur Heart J. 1988;9(suppl I):83-86.
Jones M, Schofield PM, Brooks NH, Dark JF, Moussalli H, Deiraniya AK, Lawson RAM, Rahman AN. Aortic valve replacement with combined myocardial revascularization. Br Heart J.. 1989;62:9-15.