Author + information
- Received March 10, 1999
- Revision received August 24, 1999
- Accepted December 2, 1999
- Published online March 15, 2000.
- ↵*Reprint requests and correspondence: Dr. Marc S. Penn, Department of Cell Biology, NC10, Lerner Research Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, Ohio 44195
To study the relationship between coronary angiography and in-hospital mortality in patients undergoing emergency surgery of the aorta without a history of coronary revascularization or coronary angiography before the onset of symptoms.
In the setting of acute ascending aortic dissection warranting emergency aortic repair, coronary angiography has been considered to be desirable, if not essential. The benefits of defining coronary anatomy have to be weighed against the risks of additional delay before surgical intervention.
Retrospective analysis of patient charts and the Cardiovascular Information Registry (CVIR) at the Cleveland Clinic Foundation.
We studied 122 patients who underwent emergency surgery of the aorta between January 1982 and December 1997. Overall, in-hospital mortality was 18.0%, and there was no significant difference between those who had coronary angiography on the day of surgery compared with those who had not (No: 16%, n = 81 vs. Yes: 22%, n = 41, p = 0.46). Multivariate analysis revealed that a history of myocardial infarction (MI) was the only predictor of in-hospital mortality (relative risk: 4.98 95% confidence interval: 1.48–16.75, p = 0.009); however, coronary angiography had no impact on in-hospital mortality in patients with a history of MI. Furthermore, coronary angiography did not significantly affect the incidence of coronary artery bypass grafting (CABG) during aortic surgery (17% vs. 25%, Yes vs. No). Operative reports revealed that 74% of all CABG procedures were performed because of coronary dissection, and not coronary artery disease.
These data indicate that determination of coronary anatomy may not impact on survival in patients undergoing emergency surgery of the aorta and support the concept that once diagnosed, patients should proceed as quickly as possible to surgery.
The debate regarding the need for coronary angiography before emergency repair of the thoracic aorta has intensified due to improved ability to noninvasively diagnose aortic dissection and aneurysm rupture (1–3). Arguments in favor of coronary angiography are that if a patient has critical coronary stenoses necessitating coronary artery bypass grafts (CABG): 1) it could be done while on cardiopulmonary bypass (CPB) during aortic repair, and, perhaps more compelling, 2) unknown critical coronary stenoses during CPB could increase the risk of difficult CPB weaning, perioperative myocardial infarction (MI) or death. The main argument against coronary angiography before surgical intervention is the inherent delay associated with the procedure, which increases the risk of cardiac tamponade or aortic rupture, the main causes of preoperative mortality in patients with type A aortic dissection (4). Further complicating the decision to implement coronary angiography in the preoperative management of patients with a type A aortic dissection is that coronary angiography in a patient with aortic dissection can take more time than in elective circumstances (5) and may be complicated by difficult access and the increased risk of aortic rupture and dissection due to instrumentation of the diseased aorta.
Patients with aortic dissection have a significant incidence of coronary artery disease (CAD). Creswell et al. (6) demonstrated that 35% of patients with acute type A dissection who underwent coronary angiography had at least one lesion with >50% narrowing resulting in 50% of these patients undergoing CABG at the time of emergency aortic repair. In an autopsy study, Larson and Edwards (7) found chronic CAD in 22% (27 out of 121) of patients with type I or II aortic dissection.
Kern et al. (8) and Rizzo et al. (1) have presented their respective institutional experience on the effect of defining coronary anatomy on survival in patients undergoing urgent repair of the thoracic aorta. In a series of 54 patients with either type A or type B aortic dissection diagnosed by aortography, Kern et al. (8) found a potential survival benefit of preoperative coronary angiography in those patients with type A dissection and a significant clinical history suggestive of CAD (mortality angiography [n = 5] versus no angiography [n = 11], 0% vs. 37.5%). In a study of 37 patients with acute type A dissection, Rizzo et al. (1) demonstrated no significant benefit from coronary angiography, and, in fact, found angiography to be associated with increased mortality due to aortic rupture during the delay to surgical intervention. They further showed that, in patients taken directly to surgery after noninvasive diagnosis of type A dissection (n = 19), there were no deaths due to aortic rupture or complications of CAD.
In this study, we investigated the relationship of preoperative coronary angiography and in-hospital mortality in 122 patients, without a history of open heart surgery or coronary angiography, who underwent emergency repair of the thoracic aorta between 1982 and 1997 at one tertiary institution.
The Cardiovascular Information Registry (CVIR) is a database containing clinical and laboratory information on consecutive patients undergoing cardiothoracic surgery at the Cleveland Clinic Foundation (CCF). We queried this database for all patients undergoing emergency surgery between January 1982 and December 1997. From this cohort, we identified patients with either type A aortic dissections or aortic aneurysms with impending rupture requiring emergency surgical repair. Data regarding coronary risk factors (family history of CAD at <60 years of age, current tobacco use, current therapy for diabetes, history of hypertension or history of hypercholesterolemia: total cholesterol >200 mg/dL or on lipid lowering therapy), preoperative electrocardiogram (ECG) and type of surgery and survival status were obtained.
All statistics were performed using SPSS version 8.0.0 for Windows (SPSS, Chicago, Illinois). Univariate comparisons of categorical and continuous variables were made using chi-square analysis or Fisher’s exact test and Student t test, as appropriate. Statistical significance was defined at the <0.05 level, using two-tailed distributions. Multivariate analysis was performed using logistic regression with mortality as a the dependent variable and age (per decade), greater than two cardiac risk factors, coronary angiography status and history of MI (defined as history of MI or Q waves on ECG) as covariates. For stepwise entry the probability had to be <0.05 and for removal >0.10.
We identified 2,191 patients who underwent emergency cardiothoracic surgery between January 1982 and December 1997. Of these, 232 patients had surgical procedures for which a priori knowledge of the coronary anatomy might have been beneficial. We then obtained operative reports to identify those patients who underwent surgical intervention primarily due to life threatening conditions of the aorta, including aortic dissection and impending or already ruptured aortic aneurysms (Table 1). This process excluded those patients who underwent emergency surgery for conditions such as endocarditis, CAD or iatrogenic complications from cardiac catheterization. We then eliminated all patients who had prior cardiac or aortic surgery, as well as those patient who had coronary angiography before the onset of symptoms. Our final study group consisted of 122 patients who underwent emergency repair of the aorta predominately due to acute type A dissections (Table 1). We divided this group into those who had coronary angiography the day of surgery (n = 41) and those who had not (n = 81). The demographics of these patients are displayed based on coronary angiography status (Table 2)and in-hospital death (Table 3). There were no significant differences in clinical characteristics between those who did or did not have their coronary anatomy defined preoperatively. The only significant difference in clinical characteristics between survivors and nonsurvivors was that nonsurvivors had a higher incidence of prior MI, but not Q-waves on ECG.
There was no significant difference in in-hospital mortality in those patients whose coronary anatomy had been defined before surgery (Fig. 1; Yes: 22% vs. No: 16%, p = 0.46). We determined in-hospital mortality in different subgroups (Table 4). There was no significant difference in mortality among patients classified by coronary angiography status with ≥2 risk factors for CAD or history of MI based on clinical history or ECG. Patients with a history of MI who had angiography on the day of surgery had a trend toward better survival than similar patients with undefined coronary anatomy.
Our data demonstrate no significant increase in in-hospital mortality in patients who received CABG (n = 30) at the time of emergency aortic repair compared with those who did not (Fig. 2; No CABG 16% vs. CABG 23%, p = 0.41). Furthermore, by reviewing the operative notes, we determined that 74% of CABG were performed because of involvement of the coronary artery in the aortic dissection, particularly in the group without preoperative coronary angiography (Fig. 3).
In order to identify those variables that independently predict in-hospital mortality, we performed multivariate logistic regression using only parameters available preoperatively. These included age, ≥2 coronary risk factors, history of MI (defined as history of MI or Q-waves on ECG) and whether coronary angiography had been defined. This analysis indicated that the only predictor of in-hospital mortality was a history of MI (relative risk [RR]: 4.98 95% confidence interval: 1.48 to 16.75, p = 0.009). There was no interaction between a history of MI and coronary angiography status. We further analyzed those patients with a history of MI to determine if they had an increased rate of CABG at the time of emergency surgery and if coronary angiography on the day of emergency surgery affected the CABG rate. The data in Table 5reveal that, regardless of history, most patients underwent CABG due to dissection. In fact, of the 13 patients with a history of MI, only one patient had CABG for CAD.
To determine whether some patients had unknown coronary disease at the time of the emergency repair of the aorta, which should have been surgically revascularized, we queried the database to determine how many patients have had CABG since their emergency surgery. To date, only one patient had subsequent CABG at our institution, approximately six years after emergency aortic surgery. This patient did not have coronary angiography before emergency surgery and did not have CABG at the time of emergency surgery.
Multiple modalities including transesophageal echocardiography, magnetic resonance imaging and computed tomography scans have become increasingly available over the past 15 years for the diagnosis of life-threatening pathologies of the aorta. We wanted to determine how access to these technologies has affected our preoperative coronary angiography rate in the management of these patients over time. We, therefore, divided our patient population into quintiles (n = 25, 25, 24, 24, 24) spanning the interval between 1982 and 1997 (Fig. 4A). There was a statistically significant trend towards the decreased use of coronary angiography preoperatively in these patients (first quintile: 60% vs. fifth quintile: 0%). There was no concomitant change in in-hospital mortality rates over this period of time (Fig. 4B), but the small number of events precluded adequate statistical analysis.
The risks and benefits of coronary angiography in the setting of acute life-threatening aortic pathology remain controversial. Ehrlich et al. (9) recently studied 14 preoperative risk factors that could be associated with 30-day mortality in patients undergoing surgical repair of type A aortic dissection. In their study of 109 patients, they found that increased age, renal failure, preoperative hemodynamic instability, preoperative cardiac resuscitation and lack of retrograde cerebral perfusion predicted increased 30-day mortality. Importantly, none of these patients underwent coronary angiography before surgery; however, the most common cause of perioperative death was cardiac failure (10 of 30 deaths, overall study mortality: 30%). We undertook this study to determine whether defining coronary anatomy before emergency surgical repair of the aorta had any effect on in-hospital mortality in a large cohort of patients operated on at one tertiary institution. Importantly, this study does not include those patients who died before surgical intervention, who may or may not have been waiting for coronary angiography.
We set stringent inclusion criteria for this retrospective study in order to minimize confounding variables. For this reason, we did not include patients with prior open heart surgery, prior coronary revascularization, either surgical or percutaneous, or prior coronary angiography performed before the onset of symptoms leading to the index emergency surgery. A recent study from our institution by Gillinov et al. (10) investigated the management of patients with prior open heart surgery and type A dissection. This study concludes that this subset of patients with a type A dissection infrequently have tamponade or hemodynamic collapse. Therefore, they recommend the use of coronary angiography in the preoperative management of these patients with preexisting CAD.
Among the 122 patients meeting our criteria, overall in-hospital mortality was not affected by whether coronary angiography was done preoperatively. Furthermore, univariate analysis revealed that there was no effect of coronary angiography status on mortality within subgroups of patients with unfavorable coronary risk factor profiles. There was an interesting, but statistically insignificant, trend towards increased survival in those patients with a history of MI who were studied with coronary angiography before emergency surgery.
Multivariate analysis, using a model consisting of data that would be available preoperatively (age, coronary risk factor profile, history of MI, coronary angiography status) also showed no significant relation between coronary angiography status and in-hospital mortality in this cohort. The only variable that predicted in-hospital mortality was a history of MI (RR: 4.98). To further understand the potential role a history of MI had on in-hospital mortality, we analyzed the effects of CABG and the indication of CABG in patients with or without a history of MI. Overall, concomitant CABG at the time of emergent aortic surgery had no significant effect on in-hospital mortality (p = 0.41). Furthermore, defining coronary anatomy before surgical intervention had no effect on overall CABG rate (Fig. 2). This observation is likely due to our finding that 74% of CABG performed at the time of surgical repair of the aorta is due to dissection of the coronary ostea and not to CAD.
Our study represents the largest attempt to date to evaluate the role of coronary angiography before emergency surgical repair of the aorta. Kern et al. (8) studied 54 patients undergoing urgent repair of the thoracic aorta with type A or B dissections or aneurysms. They reported an overall mortality of 24% with no statistically significant survival benefit with preoperative coronary angiography. The authors did, however, find a trend toward a survival benefit in patients who had preoperative coronary angiography with a type A dissection and a clinical history suggestive of CAD. Our study, with its significantly larger population, supports and refines these findings by demonstrating a statistically nonsignificant trend towards improved outcome in patients with a documented history of MI who had coronary angiography preoperatively. The potential implications of this trend, however, have to be tempered by our findings outlined in Table 5 that: A) only 5 of 13 patients with a history of MI had concomitant CABG, B) the majority of CABG in patients with a history of MI were performed due to coronary dissection (n = 4) and not CAD (n = 1), C) only one patient of the five with a history of MI who received coronary angiography preoperatively had CABG, and D) there was no significant difference in cardiac mortality between those patients with a history of MI who did nor did not have coronary angiography (Yes: 25%, n = 2 vs. No: 20%, n = 1).
Rizzo et al. (1) in a study of 37 patients presenting with acute dissection of the ascending aorta demonstrated less preoperative and operative mortality in those patients who, once diagnosed with an ascending aortic dissection, proceeded directly to surgical repair without coronary angiography. This study, unlike ours, included those patients who died during the time from diagnosis through coronary angiography and presentation to the operating room, which may explain why we did not observe a mortality benefit in those patients who did not have recent coronary angiography.
There were limitations to our study. We did not obtain data on left ventricular systolic function, which is known to impact survival in patients with CAD and may have a role in the increased mortality of patients with a history of MI. Also, while this study is the largest to date, the strength of statistical analysis is limited by a relatively small patient population (122), 13 of whom had a history of MI. That only 10% of our population had a history of MI is due to the fact that our inclusion criteria, in an attempt to limit confounding variables, required patients not to have had coronary angiography or coronary revascularization before the onset of symptoms.
Importantly, there undoubtedly existed a selection bias in choosing those patients who underwent coronary angiography before surgical repair. It is likely that only those patients who were relatively clinically stable were chosen for preoperative coronary angiography. This hypothesis is supported by our observation that none of the patients who underwent coronary angiography died due to complications of their aortic disease (Table 5). This selection bias makes our results all the more interesting since we did not observe a survival benefit in these patients. Thus, one could speculate that when a patient is diagnosed with life-threatening pathology of the aorta and “appears” clinically stable, emergency surgery should be immediately undertaken, since any delay may result in a worse outcome.
We concluded from this relatively large retrospective study that coronary angiography before the emergency repair of thoracic aortic dissection or rupture was not associated with improved in-hospital survival in patients without a documented history of CAD and, thus, should not be considered mandatory. In patients with a prior MI, individualized assessment of the need for preoperative coronary angiography should be undertaken. Concomitant CABG, when performed, is not associated with improved survival; furthermore, late revascularization for unidentified CAD at the time of aortic surgery is very rare.
☆ Dr. Penn is a recipient of a National Research Service Award from the National Institutes of Health (HL-09911).
- coronary artery bypass graft
- coronary artery disease
- cardiopulmonary bypass
- myocardial infarction
- relative risk
- Received March 10, 1999.
- Revision received August 24, 1999.
- Accepted December 2, 1999.
- American College of Cardiology
- Banning A.,
- Masani N.,
- Ikram S.,
- Fraser A.,
- Hall R.
- Ehrlich M.,
- Fang C.,
- Grabenwoger M.,
- Cartes-Zumelzu F.,
- Wolner E.,
- Havel M.