|
|
||||||||
Ann Thorac Surg 2005;80:2091-2097
© 2005 The Society of Thoracic Surgeons
a Cardiac Surgical Unit, Massachusetts General Hospital, Boston, Massachusetts
b Vascular Surgery Division, Massachusetts General Hospital, Boston, Massachusetts
Accepted for publication May 11, 2005.
* Address correspondence to Dr Akins, Department of Surgery, White 503, Massachusetts General Hospital, 55 Fruit St, Boston, MA02114 (Email: cakins{at}partners.org).
| Abstract |
|---|
|
|
|---|
METHODS: Hospital records of 500 consecutive patients having concomitant carotid endarterectomy and coronary artery bypass grafting between 1979 and 2001 were reviewed, allowing at least 1 year of follow-up on all patients. Long-term nonfatal complications were assessed by actual and actuarial methods.
RESULTS: Patient demographics revealed a mean age of 69 years; 74% (370 patients) were male; 75% (377 patients) presented with unstable coronary syndromes; 10% (50 patients) had an intraaortic balloon pump; and 66% (329 patients) were neurologically asymptomatic. Hospital mortality was 3.6% (18 patients). Significant multivariable predictors of hospital death were preoperative transient ischemic attack or myocardial infarction, and nonelective operation. Perioperative strokes were 4.6% (23 patients), of which 2.4% (12 patients) were ipsilateral and 2.2% (11 patients) were contralateral. Significant multivariable predictors of stroke were peripheral vascular disease and use of the right internal mammary artery. Ten-year actuarial survival was 43%. Ten-year actual versus Kaplan-Meier actuarial freedoms with 95% confidence limits from late events were myocardial infarction 87% (78% and 92%) versus 81% (75% and 87%); percutaneous coronary intervention 92% (85% and 96%) versus 89% (84% and 94%); reoperative coronary grafting 96% (89% and 99%) versus 94% (90% and 98%); total stroke 85% (77% and 91%) versus 82% (76% and 87%); ipsilateral stroke 90% (83% and 94%) versus 87% (82% and 92%); carotid endarterectomy 82% (73% and 88%) versus 75% (69% and 82%).
CONCLUSIONS: Concomitant carotid and coronary artery surgery is safe and effective, particularly in preventing ipsilateral stroke, and neutralizes the impact of unilateral carotid stenosis on early and late stroke. Actual, not actuarial, methods more accurately represent the true risk of nonfatal late events.
| Introduction |
|---|
|
|
|---|
| Material and Methods |
|---|
|
|
|---|
Perioperative myocardial infarction was defined as either new Q waves or the elevation of cardiac enzymes (creatine kinase or troponin) in association with persistent ST-segment changes or new conduction disturbances. Urgent operations were defined as operative procedures performed in patients whose accelerated symptoms prompted urgent hospital admission for evaluation and who were judged to be too unstable to discharge before operative intervention. True emergency operations were defined as procedures performed on patients whose cardiovascular instability required operative intervention either outside of normal operating hours or that displaced another patient on the surgical schedule. Nonelective operations were the sum of urgent and emergency operations.
Virtually all patients in the study were admitted to the hospital because of cardiac symptoms, and their carotid disease was diagnosed incidentally. For most of the study and universally since the early 1990s, noninvasive carotid testing was routinely performed for patients who were 65 years or older, or had neurologic symptoms, history of stroke, audible carotid bruits, evidence of extensive vascular disease, or significant left main coronary artery disease. Early in the series positive noninvasive tests suggesting severe stenosis led to conventional carotid angiography or magnetic resonance angiography. In later years most carotid operations were performed solely on the basis of the findings of noninvasive tests, mirroring the evolution of carotid surgery practice at our hospital [13].
A significant carotid stenosis was defined as one that (1) reduced the luminal diameter by 70% or more on direct carotid angiography, (2) was associated with "signal dropout" on magnetic resonance angiography, or (3) was associated with an internal carotid artery area to common carotid artery area ratio of greater than 4.0 on duplex scanning.
For patients with bilateral carotid disease, the majority had considerable difference in the severity of the disease in the two arteries. The more severe was chosen for the concomitant operation. The less severely diseased carotid was followed with serial noninvasive testing to determine the need for interval endarterectomy. Of those with severe bilateral carotid disease, 17 patients had both carotid lesions surgically corrected during the index hospitalization: 1 with bilateral carotid endarterectomies with his coronary grafting, 15 with an isolated carotid endarterectomy before the concomitant procedure, and 1 with an isolated carotid endarterectomy after the combined procedure.
Operative Technique
All patients had coronary artery bypass grafting and carotid endarterectomy of at least one carotid artery during the same anesthetic. A separate team from our Vascular Surgery Service performed the carotid endarterectomy immediately before cardiopulmonary bypass and coronary grafting. The incidence of electroencephalographic monitoring and shunting is listed in Table 1.
|
Follow-Up
Follow-up clinical information concerning survival and subsequent neurologic or cardiac events was obtained between May 23, 2002, and August 25, 2003. Follow-up was obtained by direct communication with the patients. If subsequent death, hospitalization, or cardiac or neurologic events occurred, the patient's physician or appropriate hospital record department was contacted to document findings. Of the 500 study patients, 482 survived hospitalization, and of the survivors, none were lost to follow-up; thus, follow-up was 100% complete. Mean follow-up was 5.9 years.
Statistical Analysis
Differences in preoperative characteristics and operative results between the original cohort of 200 patients and the subsequent 300 patients were assessed by Fisher's exact test or unpaired Student's t test. The statistical tests presented in Tables 1 through 3
apply only to the comparison of the initial group of 200 patients versus the current group of 300, not either of these groups versus the total combined group. The comparisons were made either with Fisher's exact test or with a
2 test.
|
|
Predictors of late events were determined by proportional hazards Cox regression models, again using forward stepping and entry and removal criteria identical to those for the logistic regression models. The BMDP 9.0 biostatistical software package (BMDP Statistical Software, Inc, Los Angeles, CA) was used for both logistic and Cox regression models, using the programs LR and P2L, respectively.
Survival, as presented in Table 4, was estimated with the use of the Kaplan-Meier method. Event-free rates for each complication ("actual rates") were estimated with the use of a Cumulative Incidence (or Competing Risk) statistical program. This is necessary to avoid the overestimation of competing nonfatal event rates that arises when deaths that occur after a nonfatal event are treated as censored in the life table of simple Kaplan-Meier treatment [14]. Both of these maximum-likelihood estimation programs were used in the NCSS 2004 statistical software package [15].
|
| Results |
|---|
|
|
|---|
From a cardiac standpoint, 264 (53%) had a prior myocardial infarction, 39 (8%) had previous coronary artery bypass grafting, and 118 (24%) had congestive heart failure. Left main coronary stenosis occurred in 211 (42%) patients, and 50 (10%) were on the intraaortic balloon pump preoperatively.
From a neurologic standpoint, 66 (13%) patients had a prior stroke, and 67 (13%) had a previous carotid endarterectomy.
Operative characteristics for the study population are recorded in Table 1. Significantly more patients in the current study group required urgent operations. Intraoperative insertion of an intraaortic balloon pump was required in 17 (3%) patients.
In-Hospital Events
The incidence of important in-hospital cardiac and neurologic events is documented in Table 3. The incidences of death, perioperative myocardial infarction, and transient ischemic attack were not statistically different between the two patient cohorts. The total stroke rate was also similar, 4.0% versus 5.0%. In the combined patient population, the rates of ipsilateral stroke versus contralateral or bilateral stroke were almost identical at 2.4% and 2.0%, respectively. For the last 300 patients, the ipsilateral stroke rate of 2.3%, although not statistically different, was less than the contralateral or bilateral stroke rate of 2.7%. Neurologic symptoms did not predict hospital stroke. Of the 23 strokes, 15 (65%) were in asymptomatic patients, which mimicked the 66% incidence of asymptomatic patients in the entire study.
Table 5 lists the significant multivariable predictors of in-hospital events: hospital death, perioperative stroke, perioperative myocardial infarction, and prolonged postoperative hospital stay (greater than 14 days).
|
|
Significant multivariable predictors of the late complications of death and stroke are listed in Table 7. Advanced age, preoperative stroke and peripheral vascular disease are significant predictors of both late complications (see Appendix).
|
| Comment |
|---|
|
|
|---|
In-hospital cardiac mortality and morbidity were essentially identical between the two groups, despite the fact that the latter cohort of 300 patients was significantly older, had more unstable coronary syndromes and more extensive coronary disease, and required more preoperative intraaortic balloon pumping and urgent operative intervention.
Although the latter cohort of patients had more bilateral carotid disease identified preoperatively, more of the patients were asymptomatic, and in-hospital neurologic morbidity was similar. Of some importance is the fact that the rate of ipsilateral stroke was virtually identical to that of bilateral or contralateral stroke. This fact suggests that our application of concomitant carotid endarterectomy and coronary artery bypass grafting has neutralized the impact of unilateral carotid stenosis as a risk factor for perioperative stroke during surgical myocardial revascularization.
The finding that use of the right internal mammary graft was associated with an increased incidence of perioperative stroke we believe to be a statistical incongruity occasionally found with multivariable analyses, because right internal mammary grafting was only used in 10 patients in the study.
The relatively poor late survival, calculated by the Kaplan-Meier method, is testimony to the advanced age of the patient population and emphasizes the deleterious impact of widespread atherosclerosis on late survival. The finding that peripheral vascular disease is a significant multivariable predictor of both late death and stroke is consistent with the report from the Northern New England Cardiovascular Disease Study Group [16].
In this study we applied a newer statistical approach to assess the late nonfatal cardiac and neurologic complications, namely actual as well as actuarial analysis. In recent cardiac surgical studies addressing valve prosthesis-related complications there has been growing enthusiasm for assessing late nonfatal complications using actual (cumulative incidence) analysis to avoid the problem of death as a competing risk factor, which is inherent in the actuarial approach [14]. Numerous studies have documented how actuarial analysis, in contrast to actual analysis, of late events can overestimate the true risk of a nonfatal late event [1719]. The data in Table 6 confirm some disparity between actual and actuarial methods of evaluating nonfatal late events in this study. We believe actual cumulative incidence approach is warranted in assessing the nonfatal late events in this study and more accurately presents the true risk of a nonfatal late complication. The excellent actual late freedom from cardiac and ipsilateral neurologic events documents the durability of this combined surgical approach.
The fact that most patients who required late carotid endarterectomy did so because of progression in the disease on the side contralateral to that addressed during the index operation is not surprising. Raman and colleagues [20] documented an increased rate of disease progression in the contralateral carotid artery after carotid endarterectomy.
Although our study does not directly address the issue of concomitant versus staged approaches, controversy persists about the relative efficacy of both approaches. Since our initial publication, Borger and colleagues [21] published a meta-analysis comparing the combined approach with either staged (initial carotid endarterectomy followed by interval coronary grafting) or reversed staged (initial coronary grafting followed by interval carotid endarterectomy) procedures. That analysis suggested that the combined approach might be associated with increased mortality and morbidity. However, there were several problems with that study design. First, meta-analysis as a statistical tool is most effective when used to combine small, randomized trials that in themselves lack sufficient statistical power to answer a question. Meta-analysis is less powerful when used to assess combined, isolated, observational reports. Of note, in the only published study with prospective randomization, Hertzer and associates [22] demonstrated an advantage to simultaneous operations versus a reversed staged approach. Second, there is a potential selection bias in the reports of staged or reversed staged procedures. Because entry criteria for virtually all reports of either staged or reversed staged approaches required the patient to have had both procedures, there must have been some patients with the staged approach who died or had sufficiently severe complications that they did not have subsequent coronary grafting, and there must have been some patients with the reversed staged approach who died or had sufficiently severe complications that they did not have subsequent carotid endarterectomy. Therefore, unless the studies of staged or reversed staged approaches were designed as intention-to-treat, the patients who did not have the subsequent planned operation would not be counted, thereby biasing the results against the combined approach.
Additionally, a recent report [23] documents the fallacy of comparing results of combined carotid endarterectomy and coronary grafting to isolated coronary artery bypass grafting because patients who require both procedures have a much higher risk profile than those having isolated myocardial revascularization. Indeed, when a risk-matched cohort of isolated coronary grafting patients was compared with patients having the combined approach to both carotid and coronary disease, the complication rates did not differ significantly.
To put our results into perspective, Table 8 lists series of concomitant carotid and coronary artery operations with 100 patients or more published since 1995. Our results are at least comparable to those from other published reports. In contrast, a multistate report [24] of community-wide outcomes of the combined operation in 226 Medicare patients documented a combined death and stroke rate of 17.7% (death, 6.6%; nonfatal stroke, 11.1%).
|
| Appendix |
|---|
| Acknowledgments |
|---|
|
|
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
R. Jin and G. L. Grunkemeier Statistical Treatment of Surgical Outcome Data Card. Surg. Adult, January 1, 2008; 3(2008): 247 - 258. [Full Text] |
||||
![]() |
M. L. Brown, T. M. Sundt III, and B. J. Gersh Indications for Revascularization Card. Surg. Adult, January 1, 2008; 3(2008): 551 - 572. [Full Text] |
||||
![]() |
C. W. Akins and R. P. Cambria Myocardial Revascularization with Carotid Artery Disease Card. Surg. Adult, January 1, 2008; 3(2008): 655 - 668. [Full Text] |
||||
![]() |
T. A. Abbruzzese and R. P. Cambria Contemporary Management of Carotid Stenosis: Carotid Endarterectomy Is Here to Stay Perspectives in Vascular Surgery and Endovascular Therapy, September 1, 2007; 19(3): 248 - 256. [Abstract] [PDF] |
||||
![]() |
G. L. Grunkemeier, R. Jin, M. J.C. Eijkemans, and J. J.M. Takkenberg Actual and Actuarial Probabilities of Competing Risks: Apples and Lemons Ann. Thorac. Surg., May 1, 2007; 83(5): 1586 - 1592. [Abstract] [Full Text] [PDF] |
||||
![]() |
O. S. Goksel Combined carotid and coronary surgery: early and late results. Ann. Thorac. Surg., October 1, 2006; 82(4): 1571 - 1572. [Full Text] [PDF] |
||||
![]() |
R. P. Cambria and C. W. Akins Reply. Ann. Thorac. Surg., October 1, 2006; 82(4): 1572 - 1572. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |