Ann Thorac Surg 2002;74:394-399
© 2002 The Society of Thoracic Surgeons
Original article: cardiovascular
Stroke after conventional versus minimally invasive coronary artery bypass
Sotiris C. Stamou, MD, PhDa,
Kathleen A. Jablonski, PhDb,
Albert J. Pfister, MDa,
Peter C. Hill, MDa,
Mercedes K.C. Dullum, MDa,
Ammar S. Bafi, MDa,
Steven W. Boyce, MDa,
Kathleen R. Petro, MDa,
Paul J. Corso, MDa*
a Section of Cardiac Surgery, Department of Surgery, Washington Hospital Center, Washington, DC, USA
b MedStar Research Institute, Washington, DC, USA
Accepted for publication March 18, 2002.
* Address reprint requests to Dr Corso, Chief Section of Cardiac Surgery, Washington Hospital Center, 106 Irving Street NW, Suite 316, South Tower, Washington, DC 20010, USA
e-mail: paul.j.corso{at}medstar.net
 |
Abstract
|
|---|
Background. Postoperative stroke is a serious complication after coronary artery bypass grafting with cardiopulmonary bypass (on-pump), and portends higher morbidity and mortality. It is unknown whether an off-pump cardiopulmonary bypass (OPCAB) approach may yield a lower stroke rate over conventional on-pump coronary artery bypass grafting.
Methods. From June 1994 to December 2000, OPCAB was performed in 2,320 patients and compared with 8,069 patients who had on-pump coronary artery bypass grafting, during the same period of time. The patients undergoing OPCAB were randomly matched to on-pump patients by propensity score. A logistic regression model was used to test the difference in the postoperative stroke rate between OPCAB and on-pump procedures controlling for the correlation between matched sets. A multiple logistic regression model predicting the risk of stroke adjusted by stroke risk factors and operation type was also computed.
Results. Matches by propensity score were found for 72% of the patients undergoing OPCAB. Patients undergoing on-pump coronary artery bypass grafting were 1.8 (95% confidence interval 1.0 to 3.1, p = 0.03) times more likely to suffer a stroke postoperatively than OPCAB patients after controlling for preoperative risk factors through matching. Independent predictors of stroke identified from the multiple logistic model included on-pump operation (versus OPCAB operation), female gender, 4 to 6 vessels grafted (versus <4 grafts), hypertension, history of previous cerebrovascular accident, carotid artery disease, chronic obstructive pulmonary disease, and depressed ejection fraction.
Conclusions. Off-pump cardiopulmonary bypass avoids the risks of cardiopulmonary bypass and atrial trauma. A substantially lower stroke rate suggests that OPCAB is a neurologically safe treatment option for revascularization.
|
Introduction
|
|---|
Major advances and refinements in surgical techniques have contributed to reductions in morbidity and mortality after coronary artery bypass grafting (CABG). However, the incidence of neurologic injury remains high. Stroke was a common complication after CABG in the 1960s (5% to 9%) [1]. Despite the improvement in surgical techniques and cardioplegic agents, along with the introduction of membrane oxygenators and in-line filtration, there is a persistent stroke rate associated with CABG, ranging from 0.8% to 5.2% [2].
Previous studies have identified risk factors of perioperative stroke after CABG with cardiopulmonary bypass (on-pump CABG) [3, 4]. Among them the most widely accepted are advanced age [4], length of cardiopulmonary bypass [5], previous cerebrovascular accident [4], carotid artery disease [4], hypertension [4], diabetes [4], and atrial fibrillation [4]. Initial evidence for lower overall postoperative cognitive dysfunction after CABG without cardiopulmonary bypass (OPCAB) was also reported [4, 6–8]. However, the absence of statistical analytical tools to match for the two types of operations significantly limits the validity of the results of previous reports [6–9].
In addition to a wide variability in the sample sizes studied, previous studies have evaluated many different variables in a nonuniform manner. To systematically investigate whether OPCAB is associated with a lower in-hospital stroke rate than the conventional on-pump CABG in a large clinical setting, we conducted this study, which includes the latest 5-year cumulative experience of CABG.
|
Material and methods
|
|---|
Patients
The computerized database of the Division of Cardiac Surgery of the Washington Hospital Center was used to identify all patients who underwent CABG in our institution between June 1994 and December 2000; 10,389 were so identified. Off-pump CABG was performed in 2,320 patients (22%) and compared with a contemporaneous group of 8,069 patients (78%) who had on-pump CABG. Only patients with six or less vessels grafted were included in the analysis. Clinical events were source-documented and adjudicated. Baseline demographics, procedural data, and perioperative outcomes were recorded and entered prospectively in a prespecified database by a dedicated data coordinating center.
Operative technique and selection criteria
Standard anesthesia and surgical techniques, extracorporeal circulation, and myocardial protection methods were used with on-pump CABG. Proximal anastomoses were performed using cross-clamp of the aorta. Very few if any T-grafts were used either with either the on-pump or OPCAB approach.
Off-pump CABG was performed using one of three surgical approaches: median sternotomy and anterior or lateral minimally invasive direct coronary artery bypass. A median sternotomy approach was favored in grafting of the right coronary artery or posterior descending branch. Main indications for anterior minimally invasive direct coronary artery bypass included isolated proximal disease of the left anterior descending or first diagonal artery, and the principal indication for lateral minimally invasive direct coronary artery bypass was bypass regrafting of the circumflex system.
Definitions
Previous stroke was defined as history of a central neurologic deficit persisting for more than 72 hours. Chronic renal insufficiency was defined as a serum creatinine value 
2.0 mg/dL. Diabetes was defined as a history of diabetes mellitus, regardless of duration of disease or need for oral agents or insulin. Recent myocardial infarction was defined as a myocardial infarction occurring within 24 hours before CABG. Prolonged ventilation was defined as the need for respiratory support for more than 24 hours. Perioperative myocardial infarction was diagnosed by the following criteria: prolonged (>20 minutes) typical angina, creatine phosphokinase-MB fraction enzyme elevation more than 40 mg/dL, and serial electrocardiogram (at least two) showing new ischemic changes. Low cardiac output syndrome was defined as the use of postoperative inotropic support for more than 24 hours. Postoperative stroke was defined as any new neurologic deficit presenting in hospital and persisting more than 72 hours.
Anesthesia and intraoperative monitoring
Routine hemodynamic, electrocardiographic, and arterial blood gas monitoring were performed during the procedures. The anesthetic protocol for OPCAB has been described elsewhere [9]. Postoperative pain control was achieved with intravenous doses of propofol (50 mg/kg) and morphine (2 mg), as needed.
Statistical analysis
Univariate comparisons of operative and postoperative characteristics were performed between the on-pump CABG and OPCAB groups using the Mantel-Haenszel 
2 test of general association unless stated otherwise. Ordinal categorical data were compared using the Cochran-Armitage test for trends. Continuous data were compared using the Wilcoxon rank test. All tests are two-sided and p values of 0.05 or less were considered significant.
Logistic regression was used to calculate the probability of being selected for on-pump CABG given a set of preoperative risk factors. These factors were identified from the literature as being associated with a stroke outcome and included age, gender, diabetes, hypertension, congestive heart failure, recent myocardial infarction, previous cerebrovascular accident, carotid artery disease, chronic renal failure on hemodialysis, chronic obstructive pulmonary disease, ejection fraction, case priority, and previous CABG [4, 10]. The number of vessels grafted and year of operation were also included. The number of vessels grafted was used as a gross indicator of severity of disease. Year of operation was included to control for variation in surgical practices over time. Model fit was evaluated using the Hosmer and Lemeshow goodness-of-fit statistic and residual analysis. The c-statistic is reported as a measure of predictive power. The presence of linear dependencies or correlation among the independent variables (multicollinearity) was checked using diagnostics from ordinary logistic regression (tolerance and the variance inflation factor) [11]. Models, which include variables that are highly correlated, produce poor estimates of their effects on the dependent variable. Propensity scores [12] or the probability of being selected for on-pump CABG were computed from these models. Patients who had OPCAB were randomly matched to those who had on-pump CABG on propensity score. By matching on propensity score, the on-pump and off-pump cases have nearly equal proportions of the preoperative variables. The matching controls for potentially confounding variables. The general estimating method [13] was used in a logistic regression to test the difference in the postoperative stroke rate between OPCAB and on-pump CABG controlling for the correlation between matched sets. For comparison purposes, a multiple logistic regression model predicting the risk of stroke adjusted by stroke risk factors and type of operation on the entire cohort was also computed. Collinearity diagnostics were carried out as in the models described. Likelihood confidence intervals were computed. Model fit was assessed using residual analysis, Hosmer and Lemeshow goodness-of-fit statistic, and the c-statistic.
|
Results
|
|---|
Unmatched sample
Univariate comparisons between type of operation and stroke risk factors are presented in Table 1.
Multivariate comparisons of the preoperative risk factors between on-pump CABG and OPCAB are shown on Table 2.
Patients in the OPCAB group were more likely to be women, have undergone a previous CABG, and be slightly younger than patients in the on-pump group. Patients in the on-pump group were more likely to have had a history of diabetes, reduced ejection fraction, and be an urgent/emergent case than patients in the OPCAB group.
View this table:
[in this window]
[in a new window]
|
Table 2. Results From a Multiple Logistic Regression Testing Preoperative Risk Factors for Association With On-Pump CABG in 10,389 Patients
|
|
Operative and postoperative patient characteristics are presented in Table 3.
In the univariate analysis on-pump CABG was associated with a higher rate of reoperation due to bleeding, low cardiac output, postoperative intraaortic balloon pump, pulmonary edema, and prolonged ventilation. Moreover, new-onset atrial fibrillation, length of stay, and operative mortality were similarly higher in the on-pump group.
View this table:
[in this window]
[in a new window]
|
Table 3. Univariate Comparisons of Operative and Postoperative Patient Characteristics Between OPCAB and On-Pump CABG
|
|
The results of the multivariate logistic regression analysis are presented in Table 4.
Independent predictors of stroke identified from the multiple logistic model, in the unmatched sample, included on-pump CABG (versus OPCAB), female gender, four to six vessels grafted (versus <4 grafts), hypertension, history of previous cerebrovascular accident, carotid artery disease, chronic obstructive pulmonary disease, and ejection fraction less than 34% (both ejection fraction <25% and ejection fraction 25% to 34% emerged as independent predictors of stroke). There was no evidence of lack of fit and there was no indication of multicollinearity among the independent variables.
View this table:
[in this window]
[in a new window]
|
Table 4. Results From a Multiple Logistic Regression Model Testing Preoperative Risk Factors for Association With Postoperative Stroke in 10,389 Patients
|
|
Matched sample
To include year of operation and control for the number of grafts, two logistic models were fit; one for patients with less than four vessels grafted, and the second with four or more vessels grafted (results not shown). Year of operation was included in both of these models. Matches by propensity score were found for 72% of the patients undergoing OPCAB. The matched sample included 3,340 patients (1,670 on each group). There was no evidence of either multicollinearity or lack of fit in the logistic regression models predicting choice of operation from preoperative risk factors. As shown in Table 5,
the on-pump CABG cases were 1.8 (95% confidence interval 1.0 to 3.0 for odds ratio, p = 0.03) times more likely to have had a postoperative stroke than the OPCAB group after controlling for preoperative risk factors through matching.
View this table:
[in this window]
[in a new window]
|
Table 5. Results From a Multiple Logistic Regression Model Testing Operation Type for Association With Postoperative Stroke in 1,670 Matches
|
|
To summarize the analysis, the unadjusted odds ratio, shows that patients undergoing on-pump CABG are 2.2 (95% CI 1.4 to 3.2) times more likely to have a stroke than off-pump patients. Table 4 shows that female gender, more than three vessels grafted, hypertension, previous cerebrovascular accident, carotid artery disease, chronic obstructive pulmonary disease, and a low ejection fraction are simultaneous risk factors for stroke. Table 2 shows that compared to patients undergoing off-pump CABG, the on-pump cases are more likely to be men, have significantly more vessels grafted, and a lower ejection fraction (hypertension, cerebrovascular accident, carotid artery disease, and chronic obstructive pulmonary disease are not significant). Adjusting for these risk factors in a multivariate analysis (Table 4) lowers the odds ratio to 1.6 (95% confidence interval 1.0 to 2.7). Adjusting for the differences in preoperative risk factors through propensity score matching raises the odds ratio to 1.8 (95% confidence interval 1.0 to 3.0) (Table 5). Adjustment through matching achieves balance and assures that the two groups will be comparable with respect to the preoperative risk factors used in the creation of the propensity score.
|
Comment
|
|---|
Neurocognitive dysfunction is a major complication after coronary revascularization. Severity may range from a stroke to a cognitive defect, or a disturbance in neuropsychologic function. As previously reported, stroke is associated with prolonged hospital stay as well as with an increased mortality and disability rate after CABG [4]. The incidence of clinically obvious stroke is reported to be between 0.8% and 5.2% and thus, between 5,000 and 35,000 new strokes occur as a result of this procedure, which makes CABG one of the largest causes of iatrogenic stroke in the United States [4].
Off-pump CABG has been used with increasing frequency in the past decade. Compared with on-pump CABG, OPCAB has been associated with decreased foreign surface/blood interactions and shear response, lower atrial fibrillation [14], stroke rates [15], and improved perioperative outcomes. The decreased organ dysfunction obtained with OPCAB has been largely ascribed to the avoidance of the systemic inflammatory response elicited by the cardiopulmonary bypass circuit [16]. Moreover, as previously reported the incidence of atrial fibrillation after OPCAB is lower than after the conventional approach, mostly related to the period of myocardial ischemia, the required atrial cannulation, and the adverse effects of cardioplegia [15]. Inadequate atrial protection has been demonstrated to be a trigger responsible for the development of atrial fibrillation in vulnerable patients [17] and therefore of thromboembolic events after on-pump CABG.
Cerebral microemboli generated during CABG with cardiopulmonary bypass might be implicated in postoperative neurologic impairment [16, 17]. The principal cause of neuropsychologic impairment after on-pump CABG was suggested to be diffuse microischemia secondary to cerebral microemboli [17]. Taylor and colleagues [17] correlated cerebral microemboli with transcranial Doppler studies and found that the greatest number of emboli were air emboli that occurred during interventions by the perfusionist (blood sampling and injections). Moody and associates [18], however, documented the presence of multiple atherosclerotic emboli lodged in the capillaries of the brain after conventional CABG. Previous investigators have also found an association between prolonged cardiopulmonary bypass time (>120 minutes) and increased risk of postoperative stroke [19]. Similarly, previous studies have demonstrated a lower release of S100B protein (a marker of brain injury) [20] and a lower incidence of high intensive transient signals in transcranial Doppler ultrasound after OPCAB versus on-pump CABG [21]. These findings emphasize the unfavorable effects of cardiopulmonary bypass on the subsequent development of stroke. In the present study, OPCAB was associated with a significantly lower stroke rate than on-pump CABG.
In addition, comorbid conditions such as previous cerebrovascular accident [5], carotid artery disease [22], hypertension [23], advanced age [24], chronic obstructive pulmonary disease, and depressed ejection fraction [5], have been found in this and previous studies to predict postoperative stroke. Interestingly, female gender in our study was associated with a higher occurrence of stroke. Female gender has been recognized as a predictor of operative mortality in a previous study [25], partially because of late referral pattern of women and possibly because of their smaller coronary arteries, which makes myocardial revascularization more demanding and the frequency of postoperative adverse events, including stroke, higher [25].
Clinical implications
Our study demonstrated an improved clinical outcome and a lower stroke rate after OPCAB versus on-pump CABG. The heightened stroke rate documented after on-pump CABG might be related to the postoperative organ dysfunction triggered by the cardiopulmonary bypass and the systemic inflammatory response it elicits (postpump syndrome). Increased capillary permeability and slight postoperative brain edema may be important corollaries of the neurologic dysfunction after on-pump CABG [23]. An amplification of these phenomena by the microemboli load has to be considered as well [23]. Avoidance of cardiopulmonary bypass is also associated with significantly lower in-hospital mortality, emphasizing the deleterious effects of the cardiopulmonary bypass in early patient survival [26, 27].
Technical improvements and better stabilization have also facilitated an increase in the rate of revascularization procedures performed on the beating heart. In our series, OPCAB became more routine overtime (on 1994 only 2% of coronary procedures were done on a beating heart, whereas the respective value for 2000 was 24%).
Limitations
A limitation of our study is that it used a retrospective single institution methodology. Despite the fact that this is among the larger analyses to date relating cardiopulmonary bypass to a higher incidence of stroke, all the limitations of a single institution study apply. Furthermore, as with our previous study on predictors of stroke after conventional on-pump CABG [4], the occurrence of clinically obvious stroke, the outcome measure of the present study might underestimate the postoperative cognitive impairment or delirium conditions. Although evaluation by an independent neurologist or brain imaging was a routine during the period of the study, there was no neuropsychologic testing that would have enabled the assessment of more subtle neurocognitive impairment, which may represent multiple territory cerebral microinfarcts [4]. Long-term efficacy and durability of minimally invasive versus conventional coronary artery operations also remains to be answered.
Another limitation was that we detected aortic arteriosclerosis by surgical palpation and not by ultrasonography; however, surgical palpation, when positive, has a high degree of specificity, as reported by Wolman and colleagues [28] in a 24-institution study. The reason we have not used ultrasonography was because it was not always available or considered to be the standard of care at the time the study was performed.
In conclusion, OPCAB is associated with a lower stroke rate as compared to the conventional on-pump CABG. However, the lack of comparative data on long-term outcomes emphasizes that these retrospective initial reports need to be validated by further studies. Thorough scientific analysis in a prospective randomized setting should ideally be performed for all surgical management paradigms.
|
References
|
|---|
- Gilman S. Cerebral disorders after open heart operations. N Engl J Med 1965;272:489-498.
- Furlan A.J., Sila C.A., Chimowitz M.I., Jones S.C. Neurologic complications related to cardiac surgery. Neurol Clin 1992;10:145-166.[Medline]
- Patel R.L., Turtle M.R., Chambers D.J., Newman S., Venn G.E. Hyperperfusion and cerebral dysfunction. Effect of differing acid-base management during cardiopulmonary bypass. Eur J Cardiothorac Surg 1993;7:457-464.[Abstract]
- Stamou S.C., Hill P.C., Dangas G., et al. Stroke after coronary artery bypass: incidence, predictors, and clinical outcome. Stroke 2001;32:1508-1513.[Abstract/Free Full Text]
- Abraham R., Karamanoukian H.L., Jajkowski M.R., et al. Does avoidance of cardiopulmonary bypass decrease the incidence of stroke in diabetics undergoing coronary surgery?. Heart Surg Forum 2001;4:1-6.
- Murkin J.M., Stump D.A. Res ipsa loquitur: protecting the brain in the new millennium, "outcomes 2000". Ann Thorac Surg 2000;69:1317-1318.[Free Full Text]
- Diegeler A., Hirsch R., Schneider F., et al. Neuromonitoring and neurocognitive outcome in off-pump versus conventional coronary bypass operation. Ann Thorac Surg 2000;69:1162-1166.[Abstract/Free Full Text]
- Ricci M., Karamanoukian H.L., Abraham R., et al. Stroke in octogenarians undergoing coronary artery surgery with and without cardiopulmonary bypass. Ann Thorac Surg 2000;69:1471-1475.[Abstract/Free Full Text]
- Resano F., Stamou S.C., Lowery R.C., Corso P.J. Coronary artery bypass grafting without cardiopulmonary bypass. Anesthetic implications. J Cardiothor Vasc Anesth 2000;14:1-8.
- Eagle K.A., Guyton R.A., Davidoff R., et al. ACC/AHA Guidelines for Coronary Artery Bypass Graft Surgery. A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Revise the 1991 Guidelines for Coronary Artery Bypass Graft Surgery). American College of Cardiology/American Heart Association. J Am Coll Cardiol 1999;34:1262-1347.[Free Full Text]
- Allison P. Logistic regression using the SAS system. Cary, NC: SAS Institute Inc, 1999.
- Rubin D.B., Thomas N. Matching using estimated propensity scores: relating theory to practice. Biometrics 1996;52:249-264.[Medline]
- Diggle P.J., Liang K.Y., Zeger S.L. The analysis of longitudinal data. New York: Oxford University Press, 1994.
- Ascione R., Caputo M., Calori G., Lloyd C.T., Underwood M.J., Angelini G.D. Predictors of atrial fibrillation after conventional and beating heart coronary surgery: a prospective, randomized study. Circulation 2000;102:1530-1535.[Abstract/Free Full Text]
- Bhasker Rao B., Van Himbergen D., Edmonds H.L., Jr, et al. Evidence for improved cerebral function after minimally invasive bypass surgery. J Cardiol Surg 1998;13:27-31.
- Selnes O.A., Goldsborough M.A., Borowicz L.M., McKhann G.M. Neurobehavioural sequelae of cardiopulmonary bypass. Lancet 1999;353:1601-1606.[Medline]
- Taylor R.L., Borger M.A., Weisel R.D., Fedorko L., Feindel C.M. Cerebral microemboli during cardiopulmonary bypass: increased emboli during perfusionist interventions. Ann Thorac Surg 1999;68:89-93.[Abstract/Free Full Text]
- Moody D.M., Bell M.A., Johnston W.E., Prough D.S. Brain microemboli during cardiac surgery or aortography. Ann Neurol 1990;28:477-486.[Medline]
- Hogue C.W., Murphy S.F., Schechtman K.B., Davila-Roman V.G. Risk factors for early or delayed stroke after cardiac surgery. Circulation 1999;100:642-647.[Abstract/Free Full Text]
- Anderson R.E., Hansson L.O., Vaage J. Release of S-100B during coronary artery bypass grafting is reduced by off-pump surgery. Ann Thorac Surg 1999;67:1721-1725.[Abstract/Free Full Text]
- Diegeler A., Hirsch R., Schneider F., et al. Neuromonitoring and neurocognitive outcome in off-pump versus conventional coronary bypass operation. Ann Thorac Surg 2000;69:1162-1166.
- Reed G.L., Singer D.E., Picard E.H., DeSanctis R.W. Stroke following coronary artery bypass surgery: a case control estimate of the risk from carotid bruits. N Engl J Med 1988;319:1246-1250.[Abstract]
- Bruer A.C., Furlan A.J., Hanson M.R., et al. Central nervous system complications of coronary bypass surgery: prospective analysis of 421 patients. Stroke 1983;14:682-687.[Abstract/Free Full Text]
- Tuman K.J., McCarthy R.J., Najafi H., Ivankovich A.D. Differential effects of advanced age on neurologic and cardiac risks of coronary artery operations. J Thorac Cardiovasc Surg 1992;104:1510-1517.[Abstract]
- Kirsch M., Guesnier L., LeBesnerais P., et al. Cardiac operations in octogenarians: perioperative risk factors for death and impaired autonomy. Ann Thorac Surg 1998;66:60-67.[Abstract/Free Full Text]
- Calafiore A.M., Di Mauro M., Contini M., et al. Myocardial revascularization with and without cardiopulmonary bypass in multivessel disease: impact of the strategy on early outcome. Ann Thorac Surg 2001;72:456-462.[Abstract/Free Full Text]
- Magee M, Jablonski KA, Stamou SC, et al. Elimination of cardiopulmonary bypass improves early survival in multivessel coronary artery bypass patients. Ann Thorac Surg (in press).
- Wolman R., Nussmeier N.A., Aggarwal A., et al. Cerebral injury after cardiac surgery: identification of a group at extraordinary risk. Stroke 1999;30:514-522.[Abstract/Free Full Text]
Top
Abstract
Introduction
