- Email: [email protected]
Increased Risk Associated With Combined Carotid Endarterectomy and Coronary Artery Bypass Graft Surgery: A Propensity-Matched Comparison With Isolated Coronary Artery Bypass Graft Surgery
Increased Risk Associated With Combined Carotid Endarterectomy and Coronary Artery Bypass Graft Surgery: A Propensity-Matched Comparison With Isolated Coronary Artery Bypass Graft Surgery Jacek B. Cywinski, MD,* Colleen Gorman Koch, MD, MS,† Leonard P. Krajewski, MD,‡ Nicholas Smedira, MD,§ Liang Li, PhD,¶ and Norman J. Starr, MD† Objective: Risk associated with combined carotid endarterectomy and coronary artery bypass graft surgery (CEA/ CABG) is controversial. The present study objective was to compare morbidity and mortality outcomes in well-matched patients who underwent combined CEA/CABG surgery with patients undergoing isolated CABG surgery with and without a history of a prior CEA. Design: This investigation was designed as a retrospective case-controlled study using data from the Cardiothoracic Anesthesia Patient Registry in a single tertiary institution. The patient population consisted of 1,698 isolated CABG surgery patients with carotid artery stenosis >40%, 708 patients who underwent an isolated CABG surgery but had a history of a prior CEA, and 272 combined CEA/CABG surgery patients who underwent surgery from January 4, 1993, through June 30, 2003. Propensity modeling techniques were used to calculate a propensity score for each patient. Greedy matching resulted in 272 propensity-matched pairs of combined CEA/CABG and isolated CABG patients (primary analysis) and 241 propensitymatched pairs of combined CEA/CABG surgery and isolated CABG surgery with previous CEA patients (secondary analysis). A Fisher exact, chi-square, Wilcoxon rank sum, and Student t test were applied appropriately to compare the propensity-matched pairs. Results: The distribution of covariates among the propensity-matched combined CEA/CABG and isolated CABG groups were similar. Among the propensity-matched pairs in the primary analysis, overall morbidity and mortality were higher in
the combined CEA/CABG group compared with the CABG group alone (overall morbidity 15% v 8.8%, p ⴝ 0.025, and mortality 5.2% v 1.1%, p ⴝ 0.007, respectively). Median intensive care unit (ICU) length of stay was longer (47 v 31 hours, p ⴝ 0.004) and hospital length of stay was longer (12 v 9 days, p < 0.001) for the combined CEA/CABG surgery compared with isolated CABG surgery, respectively. Postoperative cardiac, neurologic, serious infection, and renal morbid events were similar between the 2 groups. In the secondary analysis, the rates of mortality, overall morbidity, and neurologic morbidity were similar between the groups, whereas the median ICU and hospital length of stay were significantly longer in the combined CEA/CABG group (47.6 v 39.8 hours, p ⴝ 0.025, and 12.0 v 9.0 days, p < 0.001, respectively). Conclusions: Increased mortality and overall morbidity outcomes were found in the combined CEA/CABG group when compared with well-matched isolated CABG patients, but similar when compared with well-matched isolated CABG patients with a history of previous CEA. Patients undergoing combined CEA/CABG procedures had significantly longer ICU and hospital lengths of stay compared with patients undergoing isolated CABG procedures. © 2006 Elsevier Inc. All rights reserved.
A
with documented presence of carotid artery disease. All patients had their procedures performed in a single institution, eliminating bias related to the differences of an interinstitutional experience. In addition, the authors also analyzed outcomes in the cohort of patients undergoing isolated CABG surgery with a history of a prior CEA. This is a unique attempt to evaluate the influence of carotid artery stenosis progression on the outcomes.
LTHOUGH TREATMENT strategies for patients with isolated carotid and coronary artery disease are clearly defined, there continues to be debate with regard to management strategies for combined carotid and coronary artery disease. Current strategies include combined coronary artery bypass graft (CABG) and carotid endarterectomy (CEA) surgery during the same anesthetic or staging of the procedures with the CABG surgery performed before or after the CEA. Controversy exists as to which strategy is optimal with regard to the incidence of adverse outcomes after surgery.1-11 The investigations presented in the literature to date on the topic do not support any treatment strategy with level 1 evidence mandating a uniform therapeutic approach. The present objective was to compare morbidity and mortality outcomes in well-matched patients undergoing combined CEA/ CABG surgery with those undergoing isolated CABG surgery
From the Departments of *General Anesthesia, †Cardiothoracic Anesthesia, ‡Vascular Surgery, §Thoracic and Cardiovascular Surgery, and ¶Quantitative Health Sciences, The Cleveland Clinic Foundation, Cleveland, OH. Address reprint requests to Jacek B. Cywinski, MD, Department of General Anesthesia, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195. E-mail: [email protected] © 2006 Elsevier Inc. All rights reserved. 1053-0770/06/2006-0008$32.00/0 doi:10.1053/j.jvca.2006.01.022 796
KEY WORDS: combined carotid endarterectomy and coronary artery bypass graft surgery, coronary artery bypass graft surgery, neurologic outcomes, mortality, morbidity, propensity matching
METHODS In this retrospective cohort investigation, the patient population consisted of 1,698 patients with carotid artery disease who underwent isolated CABG surgery, 708 patients who underwent an isolated CABG surgery with a history of a prior CEA, and 272 patients who underwent a combined CEA/CABG procedure from January 4, 1993, through June 30, 2003. In the combined CEA/CABG group, carotid artery disease was significant enough to mandate surgical treatment on its own basis, whereas coronary artery disease was unfavorable for the staged procedure. All of the isolated CABG patients had carotid artery disease defined as the presence of greater than 40% stenosis and a CEA was deemed not indicated at the time of the CABG procedure. Preoperative demographic variables, comorbid conditions, laboratory values, and operative and outcome variables were collected concurrently with patient care by individuals trained in database management in the Department of Cardiothoracic Anesthesia. The 23 variables used in the logistic regression for calculation of the propensity score are listed in Table 1. Institutional review board approval was obtained from the Cleveland Clinic Foundation to perform analyses from the department registry. Postoperative outcome variables are defined in Appendix 1.
Journal of Cardiothoracic and Vascular Anesthesia, Vol 20, No 6 (December), 2006: pp 796-802
COMBINED CEA AND CABG
797
Table 1. Baseline Patient Characteristics After Matching: Primary Analysis Variables
Demographics Male gender* Age (y) Height (cm) Weight (kg) Preoperative laboratory values Blood urea nitrogen (mg/dL) Creatinine (mg/dL) Hematocrit (%) Clinical history Preoperative myocardial infarction LVF (normal v abnormal) Reoperation COPD/asthma Carotid disease ⬎40% stenosis Carotid surgery Left main coronary artery disease (⬎70%) Dialysis Heart failure Hypertension Smoking Stroke Vascular surgery or dilatations Vascular disease Diabetes Operative variables Cardiopulmonary bypass time (min) Aortic cross-clamp time (min)
Combined CABG/CEA (n ⫽ 272)
Isolated CABG (n ⫽ 272)
Mean (SD) 196 (72.1) 69.0 (8.0) 169.0 (9.4) 78.9 (14.8) Mean (SD) 21.4 (11.2) 1.3 (0.8) 38.8 (4.6) Sum (percentage) 140 (51.5) 151 (55.5) 19 (7.0) 28 (10.3) 272 (100) 0 (0) 32 (11.8) 2 (0.74) 68 (25) 213 (78.3) 200 (73.5) 43 (15.8) 17 (6.3) 77 (28.3) 117 (43) Median (interquartile range) 99.0 (46.5) 67.0 (36.5)
Mean (SD) 207 (76.1) 68.4 (8.9) 169.5 (9.5) 80.0 (14.9) Mean (SD) 20.6 (8.5) 1.3 (0.8) 39.6 (4.9) Sum (percentage) 124 (45.6) 148 (54.4) 19 (7.0) 24 (8.8) 272 (100) 0 (0) 32 (11.8) 2 (0.74) 61 (22.4) 210 (77.2) 202 (74.3) 40 (14.7) 20 (7.4) 74 (27.2) 111 (40.8) Median (interquartile range) 97.0 (48.0) 68.5 (35.5)
p Value
0.28 0.42 0.51 0.38 0.35 0.60 0.045 0.17 0.80 0.99 0.56 — — 0.99 0.99 0.48 0.76 0.85 0.72 0.61 0.77 0.60
0.78 0.44
Abbreviations: CABG, coronary artery bypass grafting; CEA, carotid endarterectomy; LVF, left-ventricular function; COPD, chronic obstructive pulmonary disease; ICU, intensive care unit. *Sum (percentage) reported.
Before propensity matching, univariate comparisons were made between the total population of combined CEA/CABG and isolated CABG patients who had not undergone a prior CEA. A Student t test, Wilcoxon rank sum, chi-square, and Fisher exact tests were applied where appropriate for group comparisons. A propensity score, in logit units, was obtained with a logistic regression that included a total of 23 covariates in the model (Table 1). The propensity score for each patient represents a summary of baseline characteristics. Greedy matching techniques were then used to select isolated CABG counterparts to combined CEA/CABG patients by choosing the patient with the nearest propensity score.12 Matching resulted in 272 propensity-matched pairs. Univariable comparisons for postoperative morbidity and mortality outcomes were made between the matched pairs with Fisher exact, chi-square, Wilcoxon rank sum, and Student t tests where appropriate. For 272 propensity-matched isolated CABG patients, summary statistics of carotid artery stenosis were calculated. The 708 patients who had undergone a CEA before the current isolated CABG procedure were propensity matched with patients undergoing CEA/CABG surgery. Greedy matching techniques resulted in 241 propensity-matched pairs. Outcome comparisons were made between the matched pairs with Fisher exact, chi-square, Wilcoxon rank sum, and Student t tests where appropriate. RESULTS
Statistically significant differences were noted in baseline characteristics between the CABG and CEA/CABG groups
before matching; patients undergoing isolated CABG surgery presented more often as a reoperation and had lower preoperative hematocrits. Patients undergoing combined CEA/CABG surgery were more often smokers. Table 1 summarizes the distribution of baseline characteristics between the propensitymatched CEA/CABG and isolated CABG groups. Summary statistics and univariate histograms of the carotid artery stenosis in the isolated CABG group after matching are presented in Figure 1. Table 2 lists the postoperative outcomes for the propensitymatched pairs. Patients who underwent combined CEA/CABG surgery had a higher incidence of overall morbidity and inhospital mortality than patients undergoing isolated CABG surgery. Median intensive care unit (ICU) and hospital lengths of stay were significantly longer in the combined CEA/CABG group compared with the matched isolated CABG patients. There were similar incidences of postoperative cardiac and respiratory morbidity, serious infection, and renal complications among the matched pairs. In addition, there were no statistically significant differences in overall neurologic morbidity, focal or global neurologic deficits, encephalopathy, or seizures between the 2 groups, although the combined CEA/ CABG group almost always had a higher neurologic morbidity than the isolated CABG group.
798
CYWINSKI ET AL
Fig 1. Summary statistics and univariate histograms of the carotid artery stenosis in the isolated CABG group after matching. LICA, left carotid artery; RICA, right carotid artery.
The secondary analysis compared patients who underwent combined CEA/CABG surgery with patients who underwent isolated CABG surgery and had a history of a prior CEA. Propensity matching resulted in a similar distribution of baseline demographics, comorbidities, and laboratory and operative variables (Table 3). Outcome comparisons are displayed in Table 4. The median ICU and hospital lengths of stay were significantly longer in the combined CEA/CABG group compared with the matched isolated CABG patients with a history of a prior CEA. The rates of mortality, overall morbidity, and neurologic morbidity were similar between the groups. Again, although statistically insignificant, there was a trend of higher neurologic complications in the combined CEA/CABG group. DISCUSSION
The optimal treatment strategy of patients with concomitant carotid and coronary artery disease is still ill defined. The major reason for that is a lack of prospective randomized trials comparing different treatment strategies. The investigations in the literature are mostly retrospective and do not include informa-
tion on the degree of carotid artery stenosis in patients undergoing isolated CABG surgery; therefore, there is a variability in examining risk associated with combined CEA/CABG surgery because of the heterogenous nature of the comparison groups. For example, there are investigations in which the comparison group is isolated CABG surgery with carotid artery disease,13 whereas others compare the combined approach with isolated CABG surgery without any degree of carotid artery disease.4 To overcome this limitation, a prospective randomized trial needs to be undertaken comparing different treatment options for these complex patients; but as of the present time, only 1 trial has been performed. Hertzer and colleagues14 showed, in a prospective randomized study, that stroke rates attributed to CABG surgery in patients randomized to combined CEA/ CABG surgery or reverse staged procedures were not statistically significant, but the early mortality rates were slightly higher in the combined CEA/CABG group.14 In the present study, the propensity matching allowed for the comparison of 2 groups (combined CEA/CABG surgery and isolated CABG surgery) matched on perioperative risk factors.
COMBINED CEA AND CABG
799
Table 2. Postoperative Outcomes for the Propensity-Matched Groups: Primary Analysis Postoperative Outcome Variables
Cardiac Postoperative myocardial infarction Low cardiac output Cardiac morbidity Neurologic Encephalopathy Focal deficit Global deficit Neurologic morbidity Seizure Pulmonary Intubation morbidity Respiratory failure Serious infection Renal Anuria Multisystem organ failure Overall morbidity Mortality Length of stay ICU (h) Hospital (d)
Combined CABG/CEA (n ⫽ 272) (%)
Isolated CABG (n ⫽ 272) (%)
Sum (percentage)
Sum (percentage)
p Value
8 (2.9) 7 (2.6) 7 (2.6)
5 (1.8) 3 (1.1) 3 (1.1)
0.40 0.20 0.20
7 (2.6) 10 (3.7) 4 (1.5) 14 (5.2) 2 (0.74)
4 (1.5) 6 (2.2) 1 (0.37) 6 (2.2) 3 (1.1)
0.36 0.31 0.37 0.068 0.99
23 (8.5) 4 (1.5) 15 (5.5)
20 (7.4) 5 (1.8) 11 (4.0)
0.63 0.99 0.42
6 (2.2) 4 (1.5) 41 (15.1) 14 (5.2) Median (interquartile range) 47.3 (65.7) 12.0 (10.0)
1 (0.37) 2 (0.74) 24 (8.8) 3 (1.1) Median (interquartile range) 30.7 (46.9) 9.0 (7.0)
0.12 0.69 0.025 0.007
0.004 ⬍ 0.001
Abbreviations: CABG, coronary artery bypass grafting; CEA, carotid endarterectomy; ICU, intensive care unit.
Although the degree of the carotid artery stenosis was not included in the matching process, the isolated CABG group had a documented stenosis of at least 40% (Fig 1), thus excluding the patients without carotid artery disease. The registry identified patients with ⱖ40% carotid artery disease as a marker for atherosclerosis of the carotid system. The authors wanted to compare the outcomes of combined CEA/CABG patients with outcomes of isolated CABG patients with proven carotid artery disease because it was speculated that atherosclerosis of the carotid system may be reflective of atherosclerosis of the aorta and attendant risks associated with surgery. Previously published studies including one by Ricotta et al15 did not report the severity of carotid stenosis in the matched patients who underwent isolated CABG surgery, thus possibly missing an important risk factor affecting the outcomes. Significant increases in overall morbidity and in-hospital mortality for patients undergoing combined CEA/CABG surgery compared with well-matched patients undergoing isolated CABG surgery were found in this single institution. Patients undergoing the combined CEA/CABG procedure also had significantly longer intensive care unit and hospital lengths of stay compared with isolated CABG patients. The occurrence of postoperative focal and global neurologic deficits, encephalopathy, and overall neurologic morbidity were statistically similar between the 2 groups; however, the combined CEA/CABG group almost always had higher neurologic morbidity rates (Table 2). These findings are in distinct contrast to results reported by Ricotta et al15 in which a similar technique of propensity matching was used to compare outcomes after combined CEA/CABG surgery and isolated CABG surgery, indi-
cating that controversy exists as to the outcomes after combined CEA/CABG procedures. Although the present study did not find statistically significant differences in neurologic outcomes between the 2 groups, mortality and overall morbidity were significantly higher in the combined CEA/CABG group. Because both groups were matched on preoperative and intraoperative variables, the results suggest that the combined procedure increases the risk of postoperative complications. The addition of CEA to a CABG procedure may not be a cause of that finding; rather, variables that were not accounted for in the matching process or an inherent risk of patients undergoing the combined procedure could play a role. In a meta-analysis of 16 studies, Borger and colleagues3 reported a significantly increased risk for the composite endpoint of stroke or death for the combined CEA/CABG procedure. Other investigators have also reported similar postoperative outcomes for the combined CEA/CABG procedure when compared with either isolated CABG or staged CEA and CABG procedures.4,13 Evagelopoulos et al4 reported similar postoperative outcomes from a retrospective analysis of 313 patients undergoing combined CEA/CABG surgery compared with isolated CABG patients without carotid artery disease. Halm et al,10 in a multicenter investigation of 2,124 patients who underwent CEA, reported a 30-day mortality or stroke rate of 10.32% for patients having the combined CEA/CABG procedure. When all patients were considered together, the risk of mortality, myocardial infarction, and stroke were directly related to the degree of patient comorbidity such that those with high comorbidity had more than twice the odds of death or stroke. These studies, however, did not clearly report the degree
800
CYWINSKI ET AL
Table 3. Baseline Patient Characteristics After Matching: Secondary Analysis Variables
Demographics Male gender* Age (y) Height (cm) Weight (kg) Preoperative laboratory values Blood urea nitrogen (mg/dL) Creatinine (mg/dL) Hematocrit (%) Clinical history Preoperative myocardial infarction LVF (normal v abnormal) Reoperation COPD/asthma Carotid disease ⬎40% stenosis Carotid surgery Left main coronary artery disease (⬎70%) Dialysis Heart failure Hypertension Smoking Stroke Vascular surgery or dilatations Vascular disease Diabetes Operative variables Cardiopulmonary bypass time (min) Aortic cross-clamp time (min)
Combined CABG/CEA (n ⫽ 241)
Isolated CABG (n ⫽ 241)
Mean (SD) 171 (71.0) 69.4 (7.7) 169.2 (9.5) 79.7 (14.9) Mean (SD) 21.1 (10.6) 1.3 (0.8) 38.7 (4.6) Sum (percentage) 127 (52.7) 130 (53.9) 19 (7.9) 26 (10.8) 241 (100) 0 (0) 28 (11.6) 2 (0.83) 59 (24.5) 192 (79.7) 178 (73.9) 42 (17.4) 17 (7.1) 74 (30.7) 99 (41.1) Median (interquartile range) 95.0 (47.0) 65.0 (34.0)
Mean (SD) 172 (71.4) 69.4 (7.9) 169.3 (9.7) 79.3 (14.5) Mean (SD) 21.1 (11.0) 1.2 (0.6) 39.2 (5.1) Sum (percentage) 128 (53.1) 129 (53.5) 24 (10.0) 25 (10.4) 231 (95.9) 241 (100) 31 (12.9) 1 (0.41) 56 (23.2) 189 (78.4) 176 (73.0) 39 (16.2) 15 (6.2) 61 (25.3) 97 (40.3) Median (interquartile range) 97.0 (49.0) 68.0 (38.0)
p Value
0.92 0.99 0.92 0.79 0.97 0.46 0.31 0.93 0.93 0.42 0.88 0.001 ⬍0.001 0.68 0.99 0.75 0.74 0.84 0.71 0.71 0.19 0.85
0.62 0.65
Abbreviations: CABG, coronary artery bypass grafting; CEA, carotid endarterectomy; LVF, left-ventricular function; COPD, chronic obstructive pulmonary disease; ICU, intensive care unit. *Sum (percentage) reported.
of carotid artery stenosis in the group of patients undergoing isolated CABG. In the present study, all patients in the isolated CABG group had carotid artery stenosis of at least 40% (Fig 1), making them more suitable for comparison with the combined CEA/CABG group. The value of ⱖ40% carotid artery disease was arbitrarily chosen as a marker for atherosclerosis of the carotid system. The higher morbidity and mortality rates of the combined procedure in the present study could be related to the addition of the CEA to the CABG procedure or to the different risk factor makeup of the patients undergoing the combined procedure; a majority of studies support the latter. Bilfinger and colleagues6 examined the observed and predicted complication rates for 84 combined CEA/CABG and 1,987 isolated CABG procedures. They reported a stroke rate of 4.76% in the combined group with ⱖ80% stenosis of the carotid artery compared with 1.7% for isolated CABG patients. Carotid endarterectomy, however, was not an independent risk factor for stroke after controlling for a number of other risk factors. Ricotta and colleagues7 compared the ability of statewide and institutional models of stroke risk after CABG surgery to predict institutionspecific results and to examine the additive stroke risk of combined CABG and CEA surgery with predictive models. An institution-specific model of stroke risk after CABG surgery was developed from 1,975 patients undergoing CABG surgery
without severe carotid disease (model I). Model I was compared with a published model (model II) derived from analysis of the same variables using New York statewide data from 1995. Predicted and observed stroke risks were compared. These formulas were applied to 154 combined CEA/CABG patients so that by determining the predicted stroke risk from CABG alone the maximal added risk could be imputed to CEA.7 When these models were applied to patients undergoing combined CEA/CABG surgery, no additional stroke risk could be attributed to the addition of CEA.7 Although their data confirm that patients who undergo combined procedures have an increased rate of postoperative stroke, their analysis suggested that the risk is primarily because of factors other than performance of a CEA. Although the etiology of stroke after CABG surgery is likely multifactorial,16,17 the literature supports the assumption that the risk of stroke in patients undergoing CABG surgery is related to the degree of carotid artery disease. In the systematic review of the literature, Naylor and colleagues2 reported that CABG patients without significant carotid artery disease (⬍50% stenosis) had an operative stroke risk of 1.8%. Unilateral (50-99%) and bilateral (50-99%) stenoses resulted in a 3.2% and 5.2% stroke risk, respectively. Patients undergoing CABG surgery with unilateral or bilateral carotid occlusion had the highest stroke risk of 7% to 12%.2
COMBINED CEA AND CABG
801
Table 4. Postoperative Outcomes for the Propensity-Matched Groups: Secondary Analysis Postoperative Outcome Variables
Cardiac Postoperative myocardial infarction Low cardiac output Cardiac morbidity Neurologic Encephalopathy Focal deficit Global deficit Neurologic morbidity Seizure Pulmonary Intubation morbidity Respiratory failure Serious infection Renal Anuria Multisystem organ failure Overall morbidity Mortality Length of stay ICU (h) Hospital (d)
Combined CABG/CEA (n ⫽ 241) (%)
Isolated CABG (n ⫽ 241) (%)
Sum (percentage)
Sum (percentage)
p Value
6 (2.5) 6 (2.5) 6 (2.5)
5 (2.1) 3 (1.2) 4 (1.7)
0.76 0.50 0.52
7 (2.9) 8 (3.3) 2 (0.83) 10 (4.2) 1 (0.41)
6 (2.5) 3 (1.2) 0 (0) 4 (1.7) 1 (0.41)
0.78 0.13 0.50 0.10 0.99
22 (9.1) 4 (1.7) 14 (5.8)
22 (9.1) 2 (0.83) 14 (5.8)
0.99 0.69 0.99
5 (2.1) 4 (1.7) 37 (15.4) 12 (5.0) Median (interquartile range) 47.6 (66.1) 12.0 (11.0)
6 (2.5) 6 (2.5) 27 (11.2) 12 (5.0) Median (interquartile range) 39.8 (46.1) 9.0 (8.0)
0.76 0.52 0.18 0.99
0.025 ⬍0.001
Abbreviations: CABG, coronary artery bypass grafting; CEA, carotid endarterectomy; ICU, intensive care unit.
There is ongoing discussion in the literature suggesting that carotid stenosis, by itself, may be a marker for other conditions, such as atherosclerotic disease of the aorta, which may also contribute to the stroke risk during CABG surgery.6 It was suggested that the majority of post-CABG strokes could not be attributed to carotid artery disease alone and that atheroembolism from the aortic arch may be the single most important cause of stroke after CABG surgery.2 The patients included in the present study had their procedures performed between 1993 and 2003, and changes in anesthetic and surgical techniques over time could have influenced mortality and morbidity rates. However, the anesthetic and surgical management have been rather consistent over the years (on-pump procedures under general anesthesia), and also patients in all groups were evenly distributed throughout the study period; therefore, changes in the management would have affected outcomes in all groups. When the combined CEA/CABG group was compared with the matched isolated CABG patients with a history of a prior CEA, the median ICU and hospital lengths of stay were significantly longer in the CEA/CABG group; however, there was no statistically significant difference in outcomes between these groups (Table 4). Interestingly, the mortality rate in the isolated CABG patients with a history of a prior CEA was the same as in the patients undergoing combined CEA/CABG surgery and much higher than in the isolated CABG patients with documented carotid artery disease. These findings might suggest that carotid artery disease is an indicator of generalized progression of atherosclerosis; hence, these patients requiring intervention for carotid stenosis before CABG surgery had more
advanced atherosclerotic disease at the time of the CABG. Another explanation of this finding could be that deaths and other serious neurologic or cardiac events after the prior CEA were not included. Thus, only the survivors and patients whose neurologic outcomes after CEA were favorable to undergo subsequent CABG surgery were included. However, even in this comparison, there was a trend of higher neurologic morbidity in the combined CEA/CABG group, although not statistically significant. A limitation of this study is that it was an observational retrospective investigation in which there may have been unaccounted or unmeasured variables that could influence the final conclusions. In this investigation, there was no statistically significant difference in the neurologic morbidity between the 2 patient groups; however, in all measured neurologic outcomes, the combined CEA/CABG group always had higher morbidity rates. This suggests that there is probably a difference between the groups, but the power of the study was not sufficient to detect it. The degree of atherosclerosis of the ascending aorta was not a measured variable in this investigation. Certainly, there has been debate over whether carotid artery stenosis, by itself, is an etiologic factor for postoperative stroke or merely a marker for atherosclerotic disease of the aorta.6 In the secondary analysis, the time interval between the CEA procedure and CABG surgery was not evaluated, potentially introducing bias by an inability to distinguish patients undergoing staged procedures from those who had a remote CEA. This study adds to the controversy of the outcomes after combined CEA/CABG procedures. As opposed to a recently published investigation using similar statistical tools,15 the present study
802
CYWINSKI ET AL
showed that even when patients are matched on preoperative risk factors, combined CEA/CABG surgery has worse outcomes than isolated CABG surgery. Most likely, retrospective studies will never fully address the safety of combined CEA/CABG procedures because of the inability to control for all perioperative risk factors. Further prospective randomized investigation is necessary to definitively examine the outcomes of different therapeutic strategies for patients with concomitant carotid stenosis and coronary
artery disease. Interestingly, this study also found that the mortality rate in the isolated CABG patients with a history of a prior CEA was the same as in the patients undergoing combined CEA/ CABG surgery and was much higher than in the isolated CABG patients with documented carotid artery disease. These findings suggest that carotid artery disease requiring intervention before CABG surgery is an indicator of increased risk after isolated CABG surgery.
APPENDIX 1: POSTOPERATIVE OUTCOME VARIABLE DEFINITIONS Mortality
Cardiac morbidity
Postoperative myocardial infarction Low cardiac output Neurologic morbidity Encephalopathy Global neurologic deficit Focal neurologic deficit Seizure Anuria Intubation morbidity Respiratory failure Multisystem organ failure Serious infection morbidity Overall morbidity
In-Hospital Death
Low cardiac output or postoperative myocardial infarction with at least one of the following: 1. Intra-aortic balloon pump (IABP) 2. Intraoperative right ventricular assist device (RVAD) or left ventricular assist device (LVAD) 3. Return to operating room for IABP, RVAD, LVAD, or ECMO placement 4. Preoperative IABP or ECMO Documented by ECG and enzymes Cardiac index ⬍1.8 L/min/m2 despite adequate fluid replacement and high-dose inotropes for ⬎4 hours Focal or global deficit in ICU or death without awakening Documented as a diagnosis by Neurology Service Symptoms global in nature (watershed infarcts). Documented by symptoms, computed tomography (CT) scan or electroencephalogram (EEG) Symptoms related to focal central lesion (aphasia, hemiparesis or hemiplegia). May or may not have central lesion documented by CT scan Documented tonic-clonic activities, requirement for anticonvulsant therapy Urine output ⬍0.5 mL/kg/8 hours. Patients on hemodialysis preoperatively are excluded Intubation time ⬎72 hours PaCO2 ⬎50 mmHg and reintubation required Failure of 3 or more organ systems Sepsis, septic shock, mediastinitis, wound infection or pneumonia. One or more morbid events
REFERENCES 1. Char D, Cuadra S, Ricotta J, et al: Combined coronary artery bypass and carotid endarterectomy: long-term results. Cardiovasc Surg 10:111-115, 2002 2. Naylor R, Cuffe RL, Rothwell PM, et al: A systematic review of outcome following synchronous carotid endarterectomy and coronary artery bypass: Influence of surgical and patient variables. Eur J Vasc Endovasc Surg 26:230-241, 2003 3. Borger M, Fremes S, Weisel R, et al: Coronary bypass and carotid endarterectomy: Does a combined approach increase risk? A metaanalysis. Ann Thorac Surg 68:14-21, 1999 4. Evagelopoulos N, Trenz M, Beckmann A, et al: Simultaneous carotid endarterectomy and coronary artery bypass grafting in 313 patients. Cardiovasc Surg 8:31-40, 2000 5. Trachiotis GD, Pfister AJ: Management strategy for simultaneous carotid endarterectomy and coronary revascularization. Ann Thorac Surg 64:1013-1018, 1997 6. Bilfinger TV, Reda H, Giron F, et al: Coronary and carotid operations under prospective standardized conditions: Incidence and outcome. Ann Thorac Surg 69:1792-1798, 2000 7. Ricotta JJ, Char DJ, Cuadra SA, et al: Modeling stroke risk after coronary artery bypass and carotid endarterectomy. Stroke 34:12121217, 2003 8. Crisby M: Risk stratification by clinical symptoms and timing of carotid endarterectomy: How could it optimize our decision-making and benefit patients with carotid stenosis? Stroke 34:2302-2303, 2003 9. Feasby TE: How appropriate is carotid endarterectomy? Stroke 34:1471-1472, 2003
10. Halm EA, Chassin MR, Tuhrim S, et al: Revisiting the appropriateness of carotid endarterectomy. Stroke 34:1464-1471, 2003 11. Naylor AR, Cuffee RL, Rothwell PM, et al: A systematic review of outcomes following staged and synchronous carotid endarterectomy and coronary artery bypass. Eur J Vasc Endovasc Surg 25:380-389, 2003 12. Reducing Bias in a Propensity Score Matched-Pair Sample Using Greedy Matching Techniques. Proceedings of the 25th Annual SAS Users Group International Conference. Cary, NC, SAS Institute, 2001 13. Zacharias A, Schwann TA, Riordan CJ, et al: Operative and 5-year outcomes of combined carotid and coronary revascularization: Review of a large contemporary experience. Ann Thorac Surg 73:491497, 2002 14. Hertzer NR, Loop FD, Beven EG, et al: Surgical staging for simultaneous coronary and carotid disease: A study including prospective randomization. J Vasc Surg 9:455-463, 1989 15. Ricotta JJ, Wall LP, Blackstone E: The influence of concurrent carotid endarterectomy on coronary bypass: A case-controlled study. J Vasc Surg 41:397-401, 2005 16. Hirotani T, Kameda T, Kumamoto T, et al: Stroke after coronary artery bypass grafting in patients with cerebrovascular disease. Ann Thorac Surg 70:1571-1576, 2000 17. Vitali E, Lanfranconi M, Bruschi G, et al: Combined surgical approach to coexistent carotid and coronary artery disease: Early and late results. Cardiovasc Surg 11:113-119, 2003
the combined CEA/CABG group compared with the CABG group alone (overall morbidity 15% v 8.8%, p ⴝ 0.025, and mortality 5.2% v 1.1%, p ⴝ 0.007, respectively). Median intensive care unit (ICU) length of stay was longer (47 v 31 hours, p ⴝ 0.004) and hospital length of stay was longer (12 v 9 days, p < 0.001) for the combined CEA/CABG surgery compared with isolated CABG surgery, respectively. Postoperative cardiac, neurologic, serious infection, and renal morbid events were similar between the 2 groups. In the secondary analysis, the rates of mortality, overall morbidity, and neurologic morbidity were similar between the groups, whereas the median ICU and hospital length of stay were significantly longer in the combined CEA/CABG group (47.6 v 39.8 hours, p ⴝ 0.025, and 12.0 v 9.0 days, p < 0.001, respectively). Conclusions: Increased mortality and overall morbidity outcomes were found in the combined CEA/CABG group when compared with well-matched isolated CABG patients, but similar when compared with well-matched isolated CABG patients with a history of previous CEA. Patients undergoing combined CEA/CABG procedures had significantly longer ICU and hospital lengths of stay compared with patients undergoing isolated CABG procedures. © 2006 Elsevier Inc. All rights reserved.
A
with documented presence of carotid artery disease. All patients had their procedures performed in a single institution, eliminating bias related to the differences of an interinstitutional experience. In addition, the authors also analyzed outcomes in the cohort of patients undergoing isolated CABG surgery with a history of a prior CEA. This is a unique attempt to evaluate the influence of carotid artery stenosis progression on the outcomes.
LTHOUGH TREATMENT strategies for patients with isolated carotid and coronary artery disease are clearly defined, there continues to be debate with regard to management strategies for combined carotid and coronary artery disease. Current strategies include combined coronary artery bypass graft (CABG) and carotid endarterectomy (CEA) surgery during the same anesthetic or staging of the procedures with the CABG surgery performed before or after the CEA. Controversy exists as to which strategy is optimal with regard to the incidence of adverse outcomes after surgery.1-11 The investigations presented in the literature to date on the topic do not support any treatment strategy with level 1 evidence mandating a uniform therapeutic approach. The present objective was to compare morbidity and mortality outcomes in well-matched patients undergoing combined CEA/ CABG surgery with those undergoing isolated CABG surgery
From the Departments of *General Anesthesia, †Cardiothoracic Anesthesia, ‡Vascular Surgery, §Thoracic and Cardiovascular Surgery, and ¶Quantitative Health Sciences, The Cleveland Clinic Foundation, Cleveland, OH. Address reprint requests to Jacek B. Cywinski, MD, Department of General Anesthesia, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195. E-mail: [email protected] © 2006 Elsevier Inc. All rights reserved. 1053-0770/06/2006-0008$32.00/0 doi:10.1053/j.jvca.2006.01.022 796
KEY WORDS: combined carotid endarterectomy and coronary artery bypass graft surgery, coronary artery bypass graft surgery, neurologic outcomes, mortality, morbidity, propensity matching
METHODS In this retrospective cohort investigation, the patient population consisted of 1,698 patients with carotid artery disease who underwent isolated CABG surgery, 708 patients who underwent an isolated CABG surgery with a history of a prior CEA, and 272 patients who underwent a combined CEA/CABG procedure from January 4, 1993, through June 30, 2003. In the combined CEA/CABG group, carotid artery disease was significant enough to mandate surgical treatment on its own basis, whereas coronary artery disease was unfavorable for the staged procedure. All of the isolated CABG patients had carotid artery disease defined as the presence of greater than 40% stenosis and a CEA was deemed not indicated at the time of the CABG procedure. Preoperative demographic variables, comorbid conditions, laboratory values, and operative and outcome variables were collected concurrently with patient care by individuals trained in database management in the Department of Cardiothoracic Anesthesia. The 23 variables used in the logistic regression for calculation of the propensity score are listed in Table 1. Institutional review board approval was obtained from the Cleveland Clinic Foundation to perform analyses from the department registry. Postoperative outcome variables are defined in Appendix 1.
Journal of Cardiothoracic and Vascular Anesthesia, Vol 20, No 6 (December), 2006: pp 796-802
COMBINED CEA AND CABG
797
Table 1. Baseline Patient Characteristics After Matching: Primary Analysis Variables
Demographics Male gender* Age (y) Height (cm) Weight (kg) Preoperative laboratory values Blood urea nitrogen (mg/dL) Creatinine (mg/dL) Hematocrit (%) Clinical history Preoperative myocardial infarction LVF (normal v abnormal) Reoperation COPD/asthma Carotid disease ⬎40% stenosis Carotid surgery Left main coronary artery disease (⬎70%) Dialysis Heart failure Hypertension Smoking Stroke Vascular surgery or dilatations Vascular disease Diabetes Operative variables Cardiopulmonary bypass time (min) Aortic cross-clamp time (min)
Combined CABG/CEA (n ⫽ 272)
Isolated CABG (n ⫽ 272)
Mean (SD) 196 (72.1) 69.0 (8.0) 169.0 (9.4) 78.9 (14.8) Mean (SD) 21.4 (11.2) 1.3 (0.8) 38.8 (4.6) Sum (percentage) 140 (51.5) 151 (55.5) 19 (7.0) 28 (10.3) 272 (100) 0 (0) 32 (11.8) 2 (0.74) 68 (25) 213 (78.3) 200 (73.5) 43 (15.8) 17 (6.3) 77 (28.3) 117 (43) Median (interquartile range) 99.0 (46.5) 67.0 (36.5)
Mean (SD) 207 (76.1) 68.4 (8.9) 169.5 (9.5) 80.0 (14.9) Mean (SD) 20.6 (8.5) 1.3 (0.8) 39.6 (4.9) Sum (percentage) 124 (45.6) 148 (54.4) 19 (7.0) 24 (8.8) 272 (100) 0 (0) 32 (11.8) 2 (0.74) 61 (22.4) 210 (77.2) 202 (74.3) 40 (14.7) 20 (7.4) 74 (27.2) 111 (40.8) Median (interquartile range) 97.0 (48.0) 68.5 (35.5)
p Value
0.28 0.42 0.51 0.38 0.35 0.60 0.045 0.17 0.80 0.99 0.56 — — 0.99 0.99 0.48 0.76 0.85 0.72 0.61 0.77 0.60
0.78 0.44
Abbreviations: CABG, coronary artery bypass grafting; CEA, carotid endarterectomy; LVF, left-ventricular function; COPD, chronic obstructive pulmonary disease; ICU, intensive care unit. *Sum (percentage) reported.
Before propensity matching, univariate comparisons were made between the total population of combined CEA/CABG and isolated CABG patients who had not undergone a prior CEA. A Student t test, Wilcoxon rank sum, chi-square, and Fisher exact tests were applied where appropriate for group comparisons. A propensity score, in logit units, was obtained with a logistic regression that included a total of 23 covariates in the model (Table 1). The propensity score for each patient represents a summary of baseline characteristics. Greedy matching techniques were then used to select isolated CABG counterparts to combined CEA/CABG patients by choosing the patient with the nearest propensity score.12 Matching resulted in 272 propensity-matched pairs. Univariable comparisons for postoperative morbidity and mortality outcomes were made between the matched pairs with Fisher exact, chi-square, Wilcoxon rank sum, and Student t tests where appropriate. For 272 propensity-matched isolated CABG patients, summary statistics of carotid artery stenosis were calculated. The 708 patients who had undergone a CEA before the current isolated CABG procedure were propensity matched with patients undergoing CEA/CABG surgery. Greedy matching techniques resulted in 241 propensity-matched pairs. Outcome comparisons were made between the matched pairs with Fisher exact, chi-square, Wilcoxon rank sum, and Student t tests where appropriate. RESULTS
Statistically significant differences were noted in baseline characteristics between the CABG and CEA/CABG groups
before matching; patients undergoing isolated CABG surgery presented more often as a reoperation and had lower preoperative hematocrits. Patients undergoing combined CEA/CABG surgery were more often smokers. Table 1 summarizes the distribution of baseline characteristics between the propensitymatched CEA/CABG and isolated CABG groups. Summary statistics and univariate histograms of the carotid artery stenosis in the isolated CABG group after matching are presented in Figure 1. Table 2 lists the postoperative outcomes for the propensitymatched pairs. Patients who underwent combined CEA/CABG surgery had a higher incidence of overall morbidity and inhospital mortality than patients undergoing isolated CABG surgery. Median intensive care unit (ICU) and hospital lengths of stay were significantly longer in the combined CEA/CABG group compared with the matched isolated CABG patients. There were similar incidences of postoperative cardiac and respiratory morbidity, serious infection, and renal complications among the matched pairs. In addition, there were no statistically significant differences in overall neurologic morbidity, focal or global neurologic deficits, encephalopathy, or seizures between the 2 groups, although the combined CEA/ CABG group almost always had a higher neurologic morbidity than the isolated CABG group.
798
CYWINSKI ET AL
Fig 1. Summary statistics and univariate histograms of the carotid artery stenosis in the isolated CABG group after matching. LICA, left carotid artery; RICA, right carotid artery.
The secondary analysis compared patients who underwent combined CEA/CABG surgery with patients who underwent isolated CABG surgery and had a history of a prior CEA. Propensity matching resulted in a similar distribution of baseline demographics, comorbidities, and laboratory and operative variables (Table 3). Outcome comparisons are displayed in Table 4. The median ICU and hospital lengths of stay were significantly longer in the combined CEA/CABG group compared with the matched isolated CABG patients with a history of a prior CEA. The rates of mortality, overall morbidity, and neurologic morbidity were similar between the groups. Again, although statistically insignificant, there was a trend of higher neurologic complications in the combined CEA/CABG group. DISCUSSION
The optimal treatment strategy of patients with concomitant carotid and coronary artery disease is still ill defined. The major reason for that is a lack of prospective randomized trials comparing different treatment strategies. The investigations in the literature are mostly retrospective and do not include informa-
tion on the degree of carotid artery stenosis in patients undergoing isolated CABG surgery; therefore, there is a variability in examining risk associated with combined CEA/CABG surgery because of the heterogenous nature of the comparison groups. For example, there are investigations in which the comparison group is isolated CABG surgery with carotid artery disease,13 whereas others compare the combined approach with isolated CABG surgery without any degree of carotid artery disease.4 To overcome this limitation, a prospective randomized trial needs to be undertaken comparing different treatment options for these complex patients; but as of the present time, only 1 trial has been performed. Hertzer and colleagues14 showed, in a prospective randomized study, that stroke rates attributed to CABG surgery in patients randomized to combined CEA/ CABG surgery or reverse staged procedures were not statistically significant, but the early mortality rates were slightly higher in the combined CEA/CABG group.14 In the present study, the propensity matching allowed for the comparison of 2 groups (combined CEA/CABG surgery and isolated CABG surgery) matched on perioperative risk factors.
COMBINED CEA AND CABG
799
Table 2. Postoperative Outcomes for the Propensity-Matched Groups: Primary Analysis Postoperative Outcome Variables
Cardiac Postoperative myocardial infarction Low cardiac output Cardiac morbidity Neurologic Encephalopathy Focal deficit Global deficit Neurologic morbidity Seizure Pulmonary Intubation morbidity Respiratory failure Serious infection Renal Anuria Multisystem organ failure Overall morbidity Mortality Length of stay ICU (h) Hospital (d)
Combined CABG/CEA (n ⫽ 272) (%)
Isolated CABG (n ⫽ 272) (%)
Sum (percentage)
Sum (percentage)
p Value
8 (2.9) 7 (2.6) 7 (2.6)
5 (1.8) 3 (1.1) 3 (1.1)
0.40 0.20 0.20
7 (2.6) 10 (3.7) 4 (1.5) 14 (5.2) 2 (0.74)
4 (1.5) 6 (2.2) 1 (0.37) 6 (2.2) 3 (1.1)
0.36 0.31 0.37 0.068 0.99
23 (8.5) 4 (1.5) 15 (5.5)
20 (7.4) 5 (1.8) 11 (4.0)
0.63 0.99 0.42
6 (2.2) 4 (1.5) 41 (15.1) 14 (5.2) Median (interquartile range) 47.3 (65.7) 12.0 (10.0)
1 (0.37) 2 (0.74) 24 (8.8) 3 (1.1) Median (interquartile range) 30.7 (46.9) 9.0 (7.0)
0.12 0.69 0.025 0.007
0.004 ⬍ 0.001
Abbreviations: CABG, coronary artery bypass grafting; CEA, carotid endarterectomy; ICU, intensive care unit.
Although the degree of the carotid artery stenosis was not included in the matching process, the isolated CABG group had a documented stenosis of at least 40% (Fig 1), thus excluding the patients without carotid artery disease. The registry identified patients with ⱖ40% carotid artery disease as a marker for atherosclerosis of the carotid system. The authors wanted to compare the outcomes of combined CEA/CABG patients with outcomes of isolated CABG patients with proven carotid artery disease because it was speculated that atherosclerosis of the carotid system may be reflective of atherosclerosis of the aorta and attendant risks associated with surgery. Previously published studies including one by Ricotta et al15 did not report the severity of carotid stenosis in the matched patients who underwent isolated CABG surgery, thus possibly missing an important risk factor affecting the outcomes. Significant increases in overall morbidity and in-hospital mortality for patients undergoing combined CEA/CABG surgery compared with well-matched patients undergoing isolated CABG surgery were found in this single institution. Patients undergoing the combined CEA/CABG procedure also had significantly longer intensive care unit and hospital lengths of stay compared with isolated CABG patients. The occurrence of postoperative focal and global neurologic deficits, encephalopathy, and overall neurologic morbidity were statistically similar between the 2 groups; however, the combined CEA/CABG group almost always had higher neurologic morbidity rates (Table 2). These findings are in distinct contrast to results reported by Ricotta et al15 in which a similar technique of propensity matching was used to compare outcomes after combined CEA/CABG surgery and isolated CABG surgery, indi-
cating that controversy exists as to the outcomes after combined CEA/CABG procedures. Although the present study did not find statistically significant differences in neurologic outcomes between the 2 groups, mortality and overall morbidity were significantly higher in the combined CEA/CABG group. Because both groups were matched on preoperative and intraoperative variables, the results suggest that the combined procedure increases the risk of postoperative complications. The addition of CEA to a CABG procedure may not be a cause of that finding; rather, variables that were not accounted for in the matching process or an inherent risk of patients undergoing the combined procedure could play a role. In a meta-analysis of 16 studies, Borger and colleagues3 reported a significantly increased risk for the composite endpoint of stroke or death for the combined CEA/CABG procedure. Other investigators have also reported similar postoperative outcomes for the combined CEA/CABG procedure when compared with either isolated CABG or staged CEA and CABG procedures.4,13 Evagelopoulos et al4 reported similar postoperative outcomes from a retrospective analysis of 313 patients undergoing combined CEA/CABG surgery compared with isolated CABG patients without carotid artery disease. Halm et al,10 in a multicenter investigation of 2,124 patients who underwent CEA, reported a 30-day mortality or stroke rate of 10.32% for patients having the combined CEA/CABG procedure. When all patients were considered together, the risk of mortality, myocardial infarction, and stroke were directly related to the degree of patient comorbidity such that those with high comorbidity had more than twice the odds of death or stroke. These studies, however, did not clearly report the degree
800
CYWINSKI ET AL
Table 3. Baseline Patient Characteristics After Matching: Secondary Analysis Variables
Demographics Male gender* Age (y) Height (cm) Weight (kg) Preoperative laboratory values Blood urea nitrogen (mg/dL) Creatinine (mg/dL) Hematocrit (%) Clinical history Preoperative myocardial infarction LVF (normal v abnormal) Reoperation COPD/asthma Carotid disease ⬎40% stenosis Carotid surgery Left main coronary artery disease (⬎70%) Dialysis Heart failure Hypertension Smoking Stroke Vascular surgery or dilatations Vascular disease Diabetes Operative variables Cardiopulmonary bypass time (min) Aortic cross-clamp time (min)
Combined CABG/CEA (n ⫽ 241)
Isolated CABG (n ⫽ 241)
Mean (SD) 171 (71.0) 69.4 (7.7) 169.2 (9.5) 79.7 (14.9) Mean (SD) 21.1 (10.6) 1.3 (0.8) 38.7 (4.6) Sum (percentage) 127 (52.7) 130 (53.9) 19 (7.9) 26 (10.8) 241 (100) 0 (0) 28 (11.6) 2 (0.83) 59 (24.5) 192 (79.7) 178 (73.9) 42 (17.4) 17 (7.1) 74 (30.7) 99 (41.1) Median (interquartile range) 95.0 (47.0) 65.0 (34.0)
Mean (SD) 172 (71.4) 69.4 (7.9) 169.3 (9.7) 79.3 (14.5) Mean (SD) 21.1 (11.0) 1.2 (0.6) 39.2 (5.1) Sum (percentage) 128 (53.1) 129 (53.5) 24 (10.0) 25 (10.4) 231 (95.9) 241 (100) 31 (12.9) 1 (0.41) 56 (23.2) 189 (78.4) 176 (73.0) 39 (16.2) 15 (6.2) 61 (25.3) 97 (40.3) Median (interquartile range) 97.0 (49.0) 68.0 (38.0)
p Value
0.92 0.99 0.92 0.79 0.97 0.46 0.31 0.93 0.93 0.42 0.88 0.001 ⬍0.001 0.68 0.99 0.75 0.74 0.84 0.71 0.71 0.19 0.85
0.62 0.65
Abbreviations: CABG, coronary artery bypass grafting; CEA, carotid endarterectomy; LVF, left-ventricular function; COPD, chronic obstructive pulmonary disease; ICU, intensive care unit. *Sum (percentage) reported.
of carotid artery stenosis in the group of patients undergoing isolated CABG. In the present study, all patients in the isolated CABG group had carotid artery stenosis of at least 40% (Fig 1), making them more suitable for comparison with the combined CEA/CABG group. The value of ⱖ40% carotid artery disease was arbitrarily chosen as a marker for atherosclerosis of the carotid system. The higher morbidity and mortality rates of the combined procedure in the present study could be related to the addition of the CEA to the CABG procedure or to the different risk factor makeup of the patients undergoing the combined procedure; a majority of studies support the latter. Bilfinger and colleagues6 examined the observed and predicted complication rates for 84 combined CEA/CABG and 1,987 isolated CABG procedures. They reported a stroke rate of 4.76% in the combined group with ⱖ80% stenosis of the carotid artery compared with 1.7% for isolated CABG patients. Carotid endarterectomy, however, was not an independent risk factor for stroke after controlling for a number of other risk factors. Ricotta and colleagues7 compared the ability of statewide and institutional models of stroke risk after CABG surgery to predict institutionspecific results and to examine the additive stroke risk of combined CABG and CEA surgery with predictive models. An institution-specific model of stroke risk after CABG surgery was developed from 1,975 patients undergoing CABG surgery
without severe carotid disease (model I). Model I was compared with a published model (model II) derived from analysis of the same variables using New York statewide data from 1995. Predicted and observed stroke risks were compared. These formulas were applied to 154 combined CEA/CABG patients so that by determining the predicted stroke risk from CABG alone the maximal added risk could be imputed to CEA.7 When these models were applied to patients undergoing combined CEA/CABG surgery, no additional stroke risk could be attributed to the addition of CEA.7 Although their data confirm that patients who undergo combined procedures have an increased rate of postoperative stroke, their analysis suggested that the risk is primarily because of factors other than performance of a CEA. Although the etiology of stroke after CABG surgery is likely multifactorial,16,17 the literature supports the assumption that the risk of stroke in patients undergoing CABG surgery is related to the degree of carotid artery disease. In the systematic review of the literature, Naylor and colleagues2 reported that CABG patients without significant carotid artery disease (⬍50% stenosis) had an operative stroke risk of 1.8%. Unilateral (50-99%) and bilateral (50-99%) stenoses resulted in a 3.2% and 5.2% stroke risk, respectively. Patients undergoing CABG surgery with unilateral or bilateral carotid occlusion had the highest stroke risk of 7% to 12%.2
COMBINED CEA AND CABG
801
Table 4. Postoperative Outcomes for the Propensity-Matched Groups: Secondary Analysis Postoperative Outcome Variables
Cardiac Postoperative myocardial infarction Low cardiac output Cardiac morbidity Neurologic Encephalopathy Focal deficit Global deficit Neurologic morbidity Seizure Pulmonary Intubation morbidity Respiratory failure Serious infection Renal Anuria Multisystem organ failure Overall morbidity Mortality Length of stay ICU (h) Hospital (d)
Combined CABG/CEA (n ⫽ 241) (%)
Isolated CABG (n ⫽ 241) (%)
Sum (percentage)
Sum (percentage)
p Value
6 (2.5) 6 (2.5) 6 (2.5)
5 (2.1) 3 (1.2) 4 (1.7)
0.76 0.50 0.52
7 (2.9) 8 (3.3) 2 (0.83) 10 (4.2) 1 (0.41)
6 (2.5) 3 (1.2) 0 (0) 4 (1.7) 1 (0.41)
0.78 0.13 0.50 0.10 0.99
22 (9.1) 4 (1.7) 14 (5.8)
22 (9.1) 2 (0.83) 14 (5.8)
0.99 0.69 0.99
5 (2.1) 4 (1.7) 37 (15.4) 12 (5.0) Median (interquartile range) 47.6 (66.1) 12.0 (11.0)
6 (2.5) 6 (2.5) 27 (11.2) 12 (5.0) Median (interquartile range) 39.8 (46.1) 9.0 (8.0)
0.76 0.52 0.18 0.99
0.025 ⬍0.001
Abbreviations: CABG, coronary artery bypass grafting; CEA, carotid endarterectomy; ICU, intensive care unit.
There is ongoing discussion in the literature suggesting that carotid stenosis, by itself, may be a marker for other conditions, such as atherosclerotic disease of the aorta, which may also contribute to the stroke risk during CABG surgery.6 It was suggested that the majority of post-CABG strokes could not be attributed to carotid artery disease alone and that atheroembolism from the aortic arch may be the single most important cause of stroke after CABG surgery.2 The patients included in the present study had their procedures performed between 1993 and 2003, and changes in anesthetic and surgical techniques over time could have influenced mortality and morbidity rates. However, the anesthetic and surgical management have been rather consistent over the years (on-pump procedures under general anesthesia), and also patients in all groups were evenly distributed throughout the study period; therefore, changes in the management would have affected outcomes in all groups. When the combined CEA/CABG group was compared with the matched isolated CABG patients with a history of a prior CEA, the median ICU and hospital lengths of stay were significantly longer in the CEA/CABG group; however, there was no statistically significant difference in outcomes between these groups (Table 4). Interestingly, the mortality rate in the isolated CABG patients with a history of a prior CEA was the same as in the patients undergoing combined CEA/CABG surgery and much higher than in the isolated CABG patients with documented carotid artery disease. These findings might suggest that carotid artery disease is an indicator of generalized progression of atherosclerosis; hence, these patients requiring intervention for carotid stenosis before CABG surgery had more
advanced atherosclerotic disease at the time of the CABG. Another explanation of this finding could be that deaths and other serious neurologic or cardiac events after the prior CEA were not included. Thus, only the survivors and patients whose neurologic outcomes after CEA were favorable to undergo subsequent CABG surgery were included. However, even in this comparison, there was a trend of higher neurologic morbidity in the combined CEA/CABG group, although not statistically significant. A limitation of this study is that it was an observational retrospective investigation in which there may have been unaccounted or unmeasured variables that could influence the final conclusions. In this investigation, there was no statistically significant difference in the neurologic morbidity between the 2 patient groups; however, in all measured neurologic outcomes, the combined CEA/CABG group always had higher morbidity rates. This suggests that there is probably a difference between the groups, but the power of the study was not sufficient to detect it. The degree of atherosclerosis of the ascending aorta was not a measured variable in this investigation. Certainly, there has been debate over whether carotid artery stenosis, by itself, is an etiologic factor for postoperative stroke or merely a marker for atherosclerotic disease of the aorta.6 In the secondary analysis, the time interval between the CEA procedure and CABG surgery was not evaluated, potentially introducing bias by an inability to distinguish patients undergoing staged procedures from those who had a remote CEA. This study adds to the controversy of the outcomes after combined CEA/CABG procedures. As opposed to a recently published investigation using similar statistical tools,15 the present study
802
CYWINSKI ET AL
showed that even when patients are matched on preoperative risk factors, combined CEA/CABG surgery has worse outcomes than isolated CABG surgery. Most likely, retrospective studies will never fully address the safety of combined CEA/CABG procedures because of the inability to control for all perioperative risk factors. Further prospective randomized investigation is necessary to definitively examine the outcomes of different therapeutic strategies for patients with concomitant carotid stenosis and coronary
artery disease. Interestingly, this study also found that the mortality rate in the isolated CABG patients with a history of a prior CEA was the same as in the patients undergoing combined CEA/ CABG surgery and was much higher than in the isolated CABG patients with documented carotid artery disease. These findings suggest that carotid artery disease requiring intervention before CABG surgery is an indicator of increased risk after isolated CABG surgery.
APPENDIX 1: POSTOPERATIVE OUTCOME VARIABLE DEFINITIONS Mortality
Cardiac morbidity
Postoperative myocardial infarction Low cardiac output Neurologic morbidity Encephalopathy Global neurologic deficit Focal neurologic deficit Seizure Anuria Intubation morbidity Respiratory failure Multisystem organ failure Serious infection morbidity Overall morbidity
In-Hospital Death
Low cardiac output or postoperative myocardial infarction with at least one of the following: 1. Intra-aortic balloon pump (IABP) 2. Intraoperative right ventricular assist device (RVAD) or left ventricular assist device (LVAD) 3. Return to operating room for IABP, RVAD, LVAD, or ECMO placement 4. Preoperative IABP or ECMO Documented by ECG and enzymes Cardiac index ⬍1.8 L/min/m2 despite adequate fluid replacement and high-dose inotropes for ⬎4 hours Focal or global deficit in ICU or death without awakening Documented as a diagnosis by Neurology Service Symptoms global in nature (watershed infarcts). Documented by symptoms, computed tomography (CT) scan or electroencephalogram (EEG) Symptoms related to focal central lesion (aphasia, hemiparesis or hemiplegia). May or may not have central lesion documented by CT scan Documented tonic-clonic activities, requirement for anticonvulsant therapy Urine output ⬍0.5 mL/kg/8 hours. Patients on hemodialysis preoperatively are excluded Intubation time ⬎72 hours PaCO2 ⬎50 mmHg and reintubation required Failure of 3 or more organ systems Sepsis, septic shock, mediastinitis, wound infection or pneumonia. One or more morbid events
REFERENCES 1. Char D, Cuadra S, Ricotta J, et al: Combined coronary artery bypass and carotid endarterectomy: long-term results. Cardiovasc Surg 10:111-115, 2002 2. Naylor R, Cuffe RL, Rothwell PM, et al: A systematic review of outcome following synchronous carotid endarterectomy and coronary artery bypass: Influence of surgical and patient variables. Eur J Vasc Endovasc Surg 26:230-241, 2003 3. Borger M, Fremes S, Weisel R, et al: Coronary bypass and carotid endarterectomy: Does a combined approach increase risk? A metaanalysis. Ann Thorac Surg 68:14-21, 1999 4. Evagelopoulos N, Trenz M, Beckmann A, et al: Simultaneous carotid endarterectomy and coronary artery bypass grafting in 313 patients. Cardiovasc Surg 8:31-40, 2000 5. Trachiotis GD, Pfister AJ: Management strategy for simultaneous carotid endarterectomy and coronary revascularization. Ann Thorac Surg 64:1013-1018, 1997 6. Bilfinger TV, Reda H, Giron F, et al: Coronary and carotid operations under prospective standardized conditions: Incidence and outcome. Ann Thorac Surg 69:1792-1798, 2000 7. Ricotta JJ, Char DJ, Cuadra SA, et al: Modeling stroke risk after coronary artery bypass and carotid endarterectomy. Stroke 34:12121217, 2003 8. Crisby M: Risk stratification by clinical symptoms and timing of carotid endarterectomy: How could it optimize our decision-making and benefit patients with carotid stenosis? Stroke 34:2302-2303, 2003 9. Feasby TE: How appropriate is carotid endarterectomy? Stroke 34:1471-1472, 2003
10. Halm EA, Chassin MR, Tuhrim S, et al: Revisiting the appropriateness of carotid endarterectomy. Stroke 34:1464-1471, 2003 11. Naylor AR, Cuffee RL, Rothwell PM, et al: A systematic review of outcomes following staged and synchronous carotid endarterectomy and coronary artery bypass. Eur J Vasc Endovasc Surg 25:380-389, 2003 12. Reducing Bias in a Propensity Score Matched-Pair Sample Using Greedy Matching Techniques. Proceedings of the 25th Annual SAS Users Group International Conference. Cary, NC, SAS Institute, 2001 13. Zacharias A, Schwann TA, Riordan CJ, et al: Operative and 5-year outcomes of combined carotid and coronary revascularization: Review of a large contemporary experience. Ann Thorac Surg 73:491497, 2002 14. Hertzer NR, Loop FD, Beven EG, et al: Surgical staging for simultaneous coronary and carotid disease: A study including prospective randomization. J Vasc Surg 9:455-463, 1989 15. Ricotta JJ, Wall LP, Blackstone E: The influence of concurrent carotid endarterectomy on coronary bypass: A case-controlled study. J Vasc Surg 41:397-401, 2005 16. Hirotani T, Kameda T, Kumamoto T, et al: Stroke after coronary artery bypass grafting in patients with cerebrovascular disease. Ann Thorac Surg 70:1571-1576, 2000 17. Vitali E, Lanfranconi M, Bruschi G, et al: Combined surgical approach to coexistent carotid and coronary artery disease: Early and late results. Cardiovasc Surg 11:113-119, 2003