Romero L, de Virgilio C, Donayre C, Stabile BE, Lewis RJ, Narahara K, Lippmann M, White R, Chang C. Trends in Cardiac Morbidity and Mortality After Endoluminal Abdominal Aortic Aneurysm Repair. Arch Surg. 2001;136(9):996-1000. doi:10.1001/archsurg.136.9.996
Copyright 2001 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2001
The adverse cardiac event rate following endoluminal abdominal aortic aneurysm (EAAA) repair has decreased as experience in performing the procedure has increased. Aneurysm complexity affects the rate of adverse cardiac events.
Design and Patients
Data from 173 consecutive patients undergoing EAAA repair from 2 successive periods were compared. There were 82 patients in the early group (group 1) and 91 patients in the later group (group 2).
Main Outcome Measures
Myocardial infarction, congestive heart failure, unstable angina, major dysrhythmias, death.
The cardiac event rate was 8.5% for group 1 vs 16.5% for group 2 (P = .16). Predictors of adverse cardiac events on multivariate analysis were the use of 4 or more graft extensions (P = .04), female sex (P = .01), and number of Eagle risk factors (P<.001). There were 2 postoperative deaths (2.4%) in group 1 and 4 (4.4%) in group 2 (P = .7).
Following EAAA repair: (1) adverse cardiac events were found to correlate with use of 4 or more graft extensions, female sex, and the number of Eagle risk factors; (2) cardiac morbidity and mortality remain significant despite greater experience and improved technology; and (3) operative mortality remains acceptably low.
ALTHOUGH there has been a trend toward reduced mortality following elective abdominal aortic aneurysm (AAA) repair during the last 4 decades, the associated cardiac morbidity remains high. Angiographic studies have confirmed that 50% of patients with AAA have severe coronary artery disease.1 Coronary artery disease is responsible for 50% to 60% of perioperative and late deaths following AAA repair.2,3 In an effort to further reduce the morbidity and mortality after AAA repair, endovascular techniques have been developed. However, we previously demonstrated no difference in adverse cardiac event rates between patients undergoing open and endoluminal AAA (EAAA) repair.4
The purpose of this study was to determine whether cardiac morbidity and mortality have decreased as a result of the increased experience in EAAA repair at our institution and to determine which perioperative factors affect the adverse cardiac event rate.
Data for consecutive patients undergoing elective EAAA repair at Harbor-UCLA Medical Center (Torrance, Calif) from July 1995 to May 1999 were reviewed retrospectively. Patients undergoing emergent operations or those with thoracoabdominal, pararenal, or suprarenal aneurysms were excluded. The patients were divided into 2 groups: group 1 consisted of patients operated on between July 1995 and June 1998 and group 2 consisted of patients operated on between July 1998 and May 1999. Complex EAAA repair was defined as requiring 4 or more graft extensions. Adverse perioperative events were defined as those occurring during the index hospital stay or within 30 days of surgery. Adverse cardiac events were defined as Q-wave and non–Q-wave myocardial infarction (MI), congestive heart failure (CHF), unstable angina, major dysrhythmias, cardiac arrest, and cardiac death. Myocardial infarction was defined as an elevation of the serum creatinine kinase MB isoenzyme level greater than or equal to 5 ng/mL and troponin level greater than or equal to 2 ng/mL, and/or new Q wave or persistent ST changes on electrocardiogram. Congestive heart failure was defined as signs or symptoms of pulmonary congestion (shortness of breath or rales); signs or symptoms of new left or right ventricular failure (cardiomegaly, jugular venous distention, peripheral edema); and abnormal chest radiography findings (vascular redistribution, interstitial, or alveolar edema). Major dysrhythmias were defined as those causing hypotension (systolic blood pressure <90 mm Hg) or those requiring monitoring in the intensive care unit. Unstable angina was defined as typical precordial chest pain consistent with ischemia, lasting 30 minutes or longer, unresponsive to nitroglycerin and rest, or a crescendo pattern of angina occurring at a lower threshold or higher frequency. Cardiac arrest and cardiac death were defined as arrest or death from a dysrhythmia or CHF caused by a cardiac complication.
Categorical variables were examined for associations by means of a χ2 or Fisher exact test where appropriate. Continuous variables were compared using the Wilcoxon rank sum test. P<.05 was considered statistically significant. Factors found to have a significant association on univariate analysis were entered into a multivariate logistic regression analysis. All statistical comparisons were performed with the SAS statistical software package (SAS Institute, Cary, NC).
One hundred seventy-three patients underwent EAAA repair (82 in group 1, 91 in group 2). The mean age was 73 years (range, 53-90 years). Ninety percent of the patients were men. The mean aneurysm size was 5.8 cm. The mean number of Eagle cardiac risk factors (age >70 years, history of MI or Q waves on electrocardiogram, diabetes, angina, ventricular ectopy requiring therapy, and CHF)5 per patient was 1.3. All patients in group 1 had Aneurx grafts (Medtronic AVE, Santa Rosa, Calif) placed. Twelve patients in group 2 had World grafts (World Medical, Sunrise, Fla) placed; the remaining patients had Aneurx grafts. In comparing groups 1 and 2, there was no difference in the mean age (P = .2), percentage of men (P = .6), mean aneurysm size P = .7), or mean number of cardiac risk factors per patient (P = .1) (Table 1). In group 1, only 1% of patients had surgery under local/regional anesthesia vs 23% in group 2 (P = .001). Mean operative time was significantly shorter for group 2 patients (180 minutes vs 226 minutes for group 1; P = .006). Group 2 patients had more complex operations, with 7 patients (7.7%) requiring 4 or more graft extensions vs 0 in group 1 (P = .01). Total length of stay was 5.3 days and 6.2 days for groups 1 and 2, respectively (P = .3).
The overall adverse cardiac event rate was 12.7%—8.5% in group 1 and 16.5% in group 2 (P = .16; odds ratio, 0.5; 95% confidence interval, 0.22-1.2). In group 1, these events included 3 patients with symptomatic arrhythmias, 2 with Q-wave MIs, and 2 with CHF. In group 2 there were 7 patients with CHF, 5 with non–Q-wave MIs, 2 with symptomatic arrhythmias, and 1 with a Q-wave MI. There were 2 perioperative deaths (2%) in group 1 and 4 (4%) in group 2 (P = .7). In group 1, both deaths were caused by multisystem organ failure. In group 2, all 4 deaths were of cardiac origin.
Patients requiring 4 or more graft extensions had an adverse cardiac event rate of 43% vs 11% for those with 3 or fewer graft extensions (P = .04). Women were found to have a 29% adverse cardiac event rate while in men the rate was 11% (P = .04). Revascularization procedures were required by 23% of the women vs 15% of the men (P = .5). The mortality rate for women and men was 6% and 3%, respectively (P = .5).
On univariate analysis, the following were associated with adverse cardiac events: the need for 4 or more graft extensions, operative time, female sex, and number of Eagle cardiac risk factors (Table 2). There were no differences between patients with and without cardiac events with respect to mean aneurysm size, estimated blood loss, and intraoperative hypotension. There were no differences between men and women with respect to the mean number of cardiac risk factors or mean operative time. On multivariate analysis, only the need for 4 or more graft extensions (P = .04), female sex (P
= .02), and number of Eagle cardiac risk factors were predictive of adverse cardiac events (P<.001) (Table 3).
This study demonstrated no difference in adverse perioperative cardiac event rates following EAAA repair performed during 2 successive periods. The adverse cardiac event rate was 8.5% in the earlier operations (group 1) and nearly twice as high (16.5%) in the later operations (group 2), although this difference was not statistically significant (P = .16). The patients were similar in age and number of Eagle cardiac risk factors. Thus, increased experience with EAAA repair did not translate into a decrease in cardiac morbidity and mortality. There are several possible explanations for this lack of difference in adverse cardiac event rates between the groups and for the trend toward a higher event rate in group 2. First, patients in group 2 may have had higher cardiac risk despite having a similar number of Eagle risk factors. The Eagle risk criteria do not incorporate such factors as recent MI or CHF or the severity of angina. Second, more patients in group 2 underwent operations under local or regional anesthesia. It is possible that this approach is more stressful than general anesthesia. We are currently investigating this issue. Furthermore, given the retrospective nature of this study, some adverse cardiac events may have been missed. Finally, the lack of difference in rate of adverse cardiac events may have been be the result of a type II statistical error.
Data on cardiac morbidity and mortality following EAAA repair is limited. May et al6 reported a 7.4% cardiac complication rate and a 2.7% cardiac mortality rate in 108 patients who underwent EAAA repair. In a smaller study, Moore et al7 noted a 2% MI rate with no cardiac deaths in 46 patients following EAAA repair. The higher adverse cardiac event rate in our series likely reflects the fact that the definition of an adverse cardiac event was more inclusive.
Although no previous studies have focused specifically on trends in adverse cardiac events following EAAA repair, May et al8 examined the effect of increased experience in performing the procedure on the overall complication rate. The authors hypothesized that the overall complication rate would improve as a result of fewer iatrogenic complications later in the learning curve. However, they reported a 64% overall complication rate in the early phase and this was only insignificantly reduced to 54% with further experience. Likewise, there was no difference in mortality rate. They attributed these similarities in perioperative morbidity and mortality rates to risks inherent to the endoluminal method. Specific cardiac morbidity was not reported.
In contradistinction, Chuter et al9 found that major perioperative morbidity (cardiac and noncardiac) decreased from 20.7% early in their experience with EAAA repair to 3.4% subsequently (P = .01). This was primarily attributed to a decrease in wound complications. The MI rate, however, was not significantly different at 6.9% in the earlier group vs 0% in the later group. Mortality rates for the 2 groups were also similar.
Our study demonstrated that adverse cardiac events correlated with the complexity of endoluminal repair. The rate of adverse cardiac events was 43% for patients requiring 4 or more graft extensions vs only 11% for those requiring 3 or fewer graft extensions (P = .04). The higher adverse cardiac event rate was not a reflection of longer operative times because the complex repairs did not take significantly longer. It is possible that patients with more complex aneurysms have more advanced coronary atherosclerosis, thus increasing their cardiac risk. Our definition of a complex repair is admittedly somewhat arbitrary, but there is no standardized way to classify aneurysm complexity. Other classification systems, such as aneurysm tortuosity and the amount of calcification or atherosclerosis, are also subjective.
We found a statistically significant sex difference with respect to cardiac events, with women having a 29% event rate compared with 11% in men. Mortality was twice as high in women (6% vs 3%), although this difference was not significant. Several explanations may account for these differences. We have noted that women with AAA seem to have more concomitant aortoiliac occlusive disease and have more difficult arterial access. The women in our study required more revascularization procedures than the men, although not significantly so. When the procedure requires a retroperitoneal incision and/or a complex arterial repair rather than a common femoral arteriotomy closure, cardiac risk may increase.10 However, women did not have significantly longer operations, nor did the mean number of Eagle cardiac risk factors differ from men. Furthermore, no women in our series required 4 or more graft extensions.
Many clinical studies have reported sex differences in the prevalence, diagnosis, treatment, and outcome for various cardiovascular diseases. Katz et al11,12 reviewed AAA repair outcomes over an 11-year period in 10 000 patients, and found that women had significantly higher mortality rates for intact (14% vs 6%) and ruptured (60% vs 50%) AAAs. They believe this sex disparity may in part be due to morphologic and genetic differences that are not yet fully understood in addition to the fact that women may not be diagnosed and treated until later in their disease. Other hypotheses are that women, in addition to being diagnosed later, are treated less aggressively than men and are underrepresented in surgical series.11 Women have also been shown to have a higher rate of operative mortality than men after coronary artery bypass grafting and heart transplantation.13,14
Our present study confirmed that the number of Eagle cardiac risk factors is predictive of adverse cardiac events after EAAA repair. In a prior study we demonstrated that the cardiac event rate increased from 0 in patients with 0 or 1 Eagle risk factor to 19.5% in patients with 2 or more Eagle risk factors.4
In conclusion, this study has shown that the operative mortality for EAAA repair remains acceptably low. We also found that the complexity of repair, female sex, and the number of Eagle risk factors are independent predictors of cardiac morbidity and mortality. Finally, despite increased experience in and improved technology for EAAA repair, cardiac morbidity and mortality remain unchanged.
Presented at the 72nd Annual Meeting of the Pacific Coast Surgical Association, Banff, Canada, February 20, 2001.
Statistical analysis was performed by Roger J. Lewis, MD, PhD.
Corresponding author: Christian de Virgilio, MD, Harbor-UCLA Medical Center, Department of Surgery, 1000 W Carson St, Torrance, CA 90509 (e-mail: email@example.com).
S. E. Wilson, MD, Orange, Calif: The 173 patients described in this presentation represent one of the larger series of endovascular repairs analyzed for cardiac morbidity. With any new device or treatment, even after release to the general medical community, complications may become apparent after experience is reached with thousands of patients, events which were not uncovered during the initial clinical trials. Thus, this paper by de Virgilio et al adds important information to our knowledge of endovascular repair. In brief, the authors report 6 deaths in 173 consecutive patients, or 3.5% mortality rate and a 12.7% overall incidence of adverse cardiac events. Inevitably this will be compared with the mortality and morbidity of patients who have had open standard operation for aortic aneurysm. For example, in the aneurysm detection and management trial (ADAM), in which over 800 patients had open surgery, the 30-day mortality was less than 2%. Recognizing that in all likelihood the authors were referred a selected group of patients biased toward higher risk, it's probable that some of these patients would not have undergone open repair. Further, since endovascular repair is assumed intuitively to have lesser risk, it would be understandable if cardiac work-up was curtailed and coronary revascularization procedures consequently omitted in very high-risk patients. Again, to compare, in the ADAM, coronary arteriograms were performed in approximately 12% of patients prior to aneurysm repair and coronary revascularization procedures performed in 2% preoperatively. Is it possible that patients undergoing endovascular repair were not as closely scrutinized for myocardial disease as would have been the case if they had had an open repair, and could this explain some of the difference in mortality rates that I have cited?
On the other hand, your patients may have reached the end of the road, with no further improvement in their cardiac status possible and still have large aneurysms needing treatment. Let me suggest that the best way to resolve this dilemma of risk stratification is for aneurysm surgeons to develop an observed over-expected outcome ratio that could be applied to patients undergoing aortic aneurysm repair whether by open or endovascular technique. This would allow us to compare outcomes in patients of similar risk status.
My first question relates to the procedure of endovascular repair itself. Is it possible that the technique of endovascular repair, where a mid-aortic balloon is quickly inflated, could cause a sudden increase in peripheral resistance and transient myocardial strain and ischemia, predisposing the patient to arrhythmias? Do you think there is anything intrinsic to graft deployment methods that could be detrimental to early postoperative myocardial function?
Secondly, because of the relatively large diameter of endovascular devices, exposure of the external iliac artery is frequently required, especially in women. This approach is more comfortable for the patient who has had a general or regional anesthetic. Could you relate myocardial complications to the magnitude of the operation necessary for access to the arterial tree? And, would this be an explanation for the higher mortality in women?
Endovascular repair is advertised almost as being an outpatient procedure. However, your patients' length of stay averaged 6.2 days, not too different from open repair. Are there any unique explanations for this somewhat long length of stay? Lastly, would the authors concede that if the mortality is close to 4% there is definitely no role for this device in the aneurysm under 5.5 cm in size? The pioneering work of White, de Virgilio, and colleagues in this area is recognized by all. We eagerly await their ongoing reports of morbidity and mortality and particularly long-term outcome on the durability of aortic endografts. The literature calls regularly for randomized trials of this technique, but in my opinion, the technology has not settled and the record of results in these early trials needs to be filled out as we have seen today before we consider randomization with standard open care.
Cornelius Olcott, MD, Stanford, Calif: I have 2 questions regarding this paper. How did you evaluate your patients from a cardiologic standpoint? Were the problems that you had with the cardiac events partially a result of poor evaluation on the part of your cardiology service, or were they things that you couldn't get around? When you are talking not so much about endovascular vs endovascular but endovascular vs open, what about the patients who have to come back for further endovascular touch-ups, replacement of further extender cuffs, other types of procedures? In our experience this is not an insignificant group. If you are going to compare those patients to an open group you have to really consider the entire operative experience, not just the 1 operation, but the times they have to come back for further procedures, either endovascular or open procedures.
Fred A. Weaver, MD, Los Angeles, Calif: I would like to understand why they chose the definition of complex as 4 extenders as opposed to 3 extenders or 2 extenders or 1 extender. Why not blood loss? Why not time? Was this definition of complexity prospectively or retrospectively derived? How would complexity be defined for endovascular prostheses that are unibody such as the Ancure graft?
Jan H. Wong, MD, Honolulu, Hawaii: There has been some evidence that beta blockade markedly reduces morbidity in these major vascular procedures. Was beta blockade uniformly utilized in these individuals?
Dr de Virgilio: The goal of this study was to see if the cardiac event rate is reduced with increased endovascular experience. We previously reported that the open technique and the endovascular technique had no difference in the cardiac event rate. To our surprise, we found that even with increased experience and shorter OR (operating room) times, the cardiac event rate has not decreased.
With respect to Dr Wilson's questions, the vast majority of patients that undergo endovascular repair are referred to us from other institutions. They are often patients who are deemed to too high risk for the open technique, whether it be from a pulmonary perspective, multiple prior laparotomies, or from a cardiac perspective. As such, I think it does need to be taken into consideration that these may not represent the typical candidates for open repair. That may be one explanation for why the event rate was so high in our series.
Another explanation for the high event rate in our series is that we did not simply define an adverse cardiac event as a Q-wave MI. To be complete, we included all cardiac events that were deemed to be significant, ie, an arrhythmia that required ICU (intensive care unit) care, CHF, etc, because we felt that this was a fair representation of important events.
With respect to what's the best method to compare patients, I agree with Dr Wilson that to date the best technique for assessing cardiac risk preoperatively is the Eagle risk factors. As he pointed out, even the Eagle risk factors have limitations. Nevertheless, we have shown that patients who have 2 or more of these cardiac risk factors have an up to 20% cardiac event rate during aortic surgery, be it open or endovascular. The finding of 2 or more of these cardiac risk factors is an important finding. I do agree though that better risk stratification is necessary in patients undergoing the endovascular approach. The Eagle risk factors don't take into consideration, for instance, how many MIs a patient has had or how recent the MI was, or how severe their angina is. These things do need to be taken into consideration.
With respect to the question of the technique of endovascular AAA, there has actually been an elegant study already in the literature looking at the hemodynamic changes during endovascular repair compared with open. It has been shown that with endovascular technique the stress with respect to changes in peripheral vascular resistance, changes in cardiac output, etc, are much less with the endovascular technique as compared with the open technique because one is occluding the aorta for a very, very brief period. I don't think that the cardiac event rate is intrinsic to the endovascular approach. I feel that it is related to the anesthetic.
With respect to the question of revascularization, we actually looked at whether the need for a revascularization increased the cardiac risk. Indeed, we have noticed parenthetically that women tend to have much smaller vessels, much more difficult access issues, but we did not find that the need for revascularization, that is, an external iliac approach, increased the cardiac risk.
Regarding length of stay, I agree that the length of stay was long in this series, but unlike other series we included the days that the patient was in the hospital preoperatively, which in many instances was 1 or 2 days. So the actual postoperative stay on average was 3½ days.
With respect to the question of mortality in the open technique and whether I would recommend the endovascular technique for a small aneurysm, I still feel that the gold standard, proven technique for a low-risk, young patient would be the open technique. The endovascular technique certainly has advantages with respect to quicker recovery time, but to date we have not demonstrated that the endovascular technique is necessarily less morbid from a cardiac perspective.
The point that Dr Olcott brought up with respect to bringing the patient back for multiple reevaluations is an important one. At our hospital, when we see patients who are indigent and are unlikely to follow up regularly, I will often recommend the open technique instead because of the logistics of having to have these patients come back on a regular basis looking for endoleaks.
On the question of cardiac evaluation, I think Dr Olcott brought up a good point that many times patients referred from outside are not undergoing the same extensive cardiac evaluation that our patients that we see in the hospital have, and I am not sure whether that makes a difference. I myself do not recommend intensive cardiac evaluation in the absence of overt cardiac symptoms.
Dr Weaver brought up the issue of how we chose 4 or more graft extensions, and I will admittedly say that this is a somewhat retrospective designation. This was chosen based on discussions with Dr Donayre and Dr White, who are my co-authors, who felt that this was a good representation of a complex repair. Since we did not study the unibody, Ancure device in this study, we will have to come up with a different definition for complex repair in that setting. The problem is that when you look at tortuosity of the vessel or look at the amount of calcification of the vessel, which are other ways of assessing how complex a repair is going to be, those criteria are themselves also somewhat contrived.
Finally, Dr Wong, you mentioned the question of beta blockade. We are strong believers in the use of beta blockade as the literature has demonstrated that this does in fact lower the perioperative cardiac morbidity and mortality, and the vast majority of our patients are placed on beta blockade, although we did not specifically address this question.