| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2008;136:29-36
© 2008 The American Association for Thoracic Surgery
Surgery for Acquired Cardiovascular Disease |
Department of Molecular and Clinical Medicine/Cardiothoracic Surgery, Sahlgrenska Academy at Göteborg University, Göteborg, Sweden.
Received for publication January 21, 2007; revisions received July 5, 2007; accepted for publication July 10, 2007. * Address for reprints: Mats J. Synnergren, MD, PhD, Children's Heart Centre, The Queen Silvia Children's Hospital, 416 85 Göteborg, Sweden. (Email: mats.synnergren{at}vgregion.se).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
Methods: A total of 9408 patients undergoing coronary artery bypass grafting, 8461 on pump and 947 off pump, operated on between 1995 and 2004 were included in the study. Adjusted hazard function for long-term mortality was estimated with Poisson regression analysis in a model that included variables reflecting completeness of revascularization, operative method (on/off pump), and background risk factors for death.
Results: Mean follow-up after surgical intervention for survivors was 5.0 ± 2.8 years (range, 0.5–10.5 years), with a total follow-up of 45,076 patient years. Leaving 1 diseased vascular segment without a bypass graft in 2- or 3-vessel disease did not increase the hazard ratio for death in comparison with complete revascularization (hazard ratio, 1.05; 95% confidence interval, 0.87–1.27; P = .60). In contrast, leaving 2 vascular segments without a bypass graft in 3-vessel disease was associated with an increased hazard ratio for death (hazard ratio, 1.82; 95% confidence interval, 1.15–2.85; P = .01). Incomplete revascularization was more common in the off-pump group (P < .001) in our study. If adjusting for incomplete revascularization, there was no significant influence of the use of on/off-pump techniques on the hazard ratio for death (hazard ratio, 1.08; 95% confidence interval, 0.82–1.40; P = .57).
Conclusions: Incomplete revascularization of patients with 3-vessel disease is an independent risk factor for increased long-term mortality after coronary artery bypass grafting. In contrast, the use of on- or off-pump techniques had no significant effect on survival after adjusting for incomplete revascularization.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
|
|
One of the major objectives of coronary artery bypass grafting (CABG) is to reduce the risk of fatal myocardial infarction. According to past studies, this is achieved in patients with coronary disease that affects a large part of the heart, most commonly patients with 3-vessel disease.1-3
Conversely, it was shown in early studies that survival benefit was reduced if revascularization was incomplete (ie, when 
1 of the major diseased vessels had not been bypassed).4 The necessity of complete revascularization has been questioned, along with the introduction of percutaneous coronary intervention (PCI) and off-pump surgery. These techniques are less traumatic, in particular PCI, but complete revascularization can be more difficult to achieve. One argument for questioning the concept of completeness is that it is grounded on the outcome of operations performed more than 20 years ago, and because medication for secondary prevention has changed so much during the years, prior experience could be obsolete.
The purpose of the present study was to analyze the influence, if any, of incomplete revascularization and on/off-pump techniques on long-term mortality after CABG in a prospective consecutive series of 9408 patients operated on between 1995 and 2004.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
There was no significant difference in the proportion of included and excluded patients between the on-pump and off-pump groups. There was no systematic registration of the reason for incomplete revascularization.
Definitions
A stenosis of more than 50% on preoperative angiographic analysis was considered significant, with the need for revascularization.5 Coronary stenoses were classified as located in any of 3 segments: the left anterior descending coronary artery (LAD) segment (LAD and diagonal arteries), the circumflex artery segment (intermediate coronary arteries and obtuse marginal branches), and the right coronary artery segment (posterior descending artery [regardless of dominance] and posterior lateral artery). Complete revascularization was defined as when at least 1 bypass had been performed on each segment with at least 1 stenosis. Incomplete revascularization was defined as when 1 or more diseased segments did not receive any bypass. Patients scheduled for reoperation were considered to have 3-vessel disease if they at any time presented with a stenosis in each of the 3 segments. Patients scheduled for reoperation with functioning grafts were considered to need a graft to a segment if there was an additional stenosis in that segment within reach for revascularization. Early mortality was defined as death for any reason within 30 days of the operation, and overall mortality was defined as death for any reason during the study period. The Cleveland Clinic risk score was used for perioperative mortality and morbidity risk stratification.6 It includes the need for emergency CABG, renal dysfunction, left ventricular dysfunction, planned reoperation, age, previous vascular surgery, chronic lung disease, anemia, body weight, diabetes mellitus, and cerebrovascular disease, as well as including whether concomitant surgical intervention of mitral valve insufficiency or aortic valve stenosis was indicated. In patients with stable angina pectoris, the severity of symptoms was classified by using the Canadian Cardiovascular Society score.7 Unstable angina was considered as new onset of severe angina, impairment of stable angina, angina at rest, or angina after myocardial infarction8 in combination with the need for hospitalization until bypass surgery. Diabetes includes types I and II.
On/off-pump Selection
The choice of the on-pump or off-pump technique was the surgeon's preference. Factors that generally favored the off-pump technique were angiographically well-sized noncalcified arteries with epicardial location, extensive aortic calcifications, and relative contraindications for cardiopulmonary bypass (previous stroke). Occasionally, multiple comorbidity (diabetes, renal, and respiratory disease) was the indication for an off-pump operation.
End Point
The primary end point was all-cause mortality obtained from the population register administered by the Swedish National Tax Board. Specific cause of death was not available.
Statistical Analysis
The purpose of this study was to determine whether long-term survival after CABG was related to completeness of revascularization, use of either off-pump or on-pump surgery, or both. This issue was approached first through a comparison of survival curves for complete revascularization versus incomplete revascularization to 1 or 2 segments and off-pump CABG versus on-pump CABG. Cumulative survival was calculated according to the Kaplan–Meier method, followed by the log-rank test. This was followed by testing for differences in background variables between the on-pump and off-pump cohorts with the Student t test for continuous variables and the 
2 test for categorical variables. Then, as the third step, a multivariate Poisson regression model was used to adjust the hazard ratio for death as a function of completeness of revascularization and on-pump versus off-pump surgery for background variables. The death hazard function after surgical treatment was estimated by use of Poisson regression.9 The hazard function was of the form exp(β0 + β1
·
x
1 + 
+ βk
·
xk), where x
1, ..., and x
k were values of the variables studied, and the β values were coefficients. The hazard function was continuous as a function of time since surgical intervention, current age, and all other continuous variables included in the model. A stepwise procedure was applied to achieve a multivariate model. Each variable was first studied together with time since surgical intervention, current age, and the interaction between the variable and a special function of time since surgical intervention. The special function was equal to the minimum of time and 0.5 years. The variables of significant importance were then included in a forward multivariate procedure. The hazard ratio can be calculated from the hazard function as the exp(Variable x β) (yes = 1 and no = 0 for categorical variables and the difference in the numeric value of the variable for continuous variables). Poisson regression is an alternative to the Cox regression model because it can handle cumulative exposures and other time-dependent covariates. It is also possible to estimate the effects of hypothetic changes of the importance in the included variables. Results are presented as means ± standard deviation.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
The mean age at operation was 66 ± 9 years (range, 29–90 years), with 78% male patients. There was a significantly higher percentage of patients with 3-vessel disease among the on-pump group but a significantly lower percentage of patients in whom the index operation was a planned reoperation compared with the off-pump group. There was also a significantly higher percentage with hypertension, previous myocardial infarction, and older age in the on-pump group but a significantly lower percentage with female sex compared with the off-pump group. The patients in the on-pump group had significantly lower ejection fractions compared with the patients in the off-pump group (Table 2). Incomplete revascularization was more frequent with off-pump surgery (Table 2).
|
|
Mortality
One thousand two hundred twelve patients died during the study period. The early mortality was 1.6% (on pump: 1.6% vs off pump: 1.1%, P = .28), and the overall mortality during follow-up was 12.9% (on pump: 13.3% vs off pump: 9.4%, P < .001).
Unadjusted Survival Curves
Survival curves, presented as Kaplan–Meier survival curves, demonstrate inferior survival in patients with 2 unbypassed vascular segments in comparison with patients with complete revascularization (Figure 1). The use of on- or off-pump techniques was associated with equal survival curves in the entire patient cohort (Figure 2), whereas in patients with 3-vessel disease, survival was inferior after off-pump procedures (Figure 3).
|
|
|
|
| (1) |
|
| (2) |
|
|
|
On/off-pump techniques
The use of on/off-pump techniques did not significantly influence the hazard ratio for death when incomplete revascularization was included as a background risk variable (hazard ratio, 1.08; 95% CI, 0.82–1.40; P = .57; Figure 6).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
Survival After CABG
In the early days of CABG, during the late 1970s and early 1980s, 3 major randomized studies compared CABG with pharmacologic treatment.1-3 It was found that selected patients with a large part of the heart affected by coronary lesions and with increased risk of premature death fared better with CABG than with pharmacologic treatment. Thus current guidelines, such as Evidence-based cardiology (Salim Yusuf, BMJ Books, 2003),10 state that patients with left main lesions, 3-vessel disease, or proximal LAD lesions have a survival benefit with CABG. However, the survival benefit of CABG is sometimes questioned. It is argued that the scientific evidence is old and newer pharmacologic agents might have reduced the risk of fatal myocardial infarction to such a degree that CABG has no further effect.
Incomplete Revascularization
The same argument that puts the survival benefit from CABG into question also applies to the necessity of complete revascularization. The studies that show that incompleteness is associated with a greater risk of premature death4,11,12 have recruited all or most of their patients with the primary operations performed more than 20 years ago. Further doubt on the completeness concept is caused by reports from the Bypass Angioplasty Revascularization Investigation13 and the Arterial Revascularization Therapies Study (ARTS),15 which found no effect on survival by incomplete revascularization. The validity of these studies is, however, compromised by their patient mix. Both studies were designed to compare CABG with PCI. Therefore all patients had to be suitable for either technique, which is the reason why both studies are dominated by patients at low and moderate risk. In these patients no survival benefit can be expected, nor can incompleteness of revascularization be expected to affect survival.
In our study patients with 3-vessel disease who only received 1 graft, which in the majority of cases was to the LAD, had a worse outcome than those who had 2 or more regions bypassed. The validity of our finding could be challenged. There were only 124 patients with 3-vessel disease and only 1 bypass graft of a total of 9408 patients. Many other influences on survival could have been lost in our analysis of such a small group. This concern would implicate that other risk factors than those listed in Table 1 (and with no relationship to the Table 1 variables) explained both incomplete revascularization and increased risk of death during follow-up. This is possible, but it is in our view not probable that any unknown variable, with no association to conventional risk variables for death, could explain both incomplete revascularization and increased mortality during follow-up. Furthermore, our finding is conceivable and is in line with previous data.4,11,12 The Coronary Artery Surgery Study (CASS),4 which enrolled patients more than 20 years earlier, made a similar observation. In their study patients with mild angina had a worse survival when only 1 vessel in 3-vessel disease was bypassed, whereas 1 omitted bypass did not affect outcome. In contrast, already 1 omitted graft in patients with severe angina or poor left ventricular function increased the risk of fatal outcome. Differences in statistical design between our study and the CASS make comparisons hazardous. With this reservation, it appears that our study population, in which the majority of patients had severe angina, had a similar outcome as the patients with "mild angina" in the CASS. If so, this could reflect more efficient secondary prevention today than 20 years ago.
There is one crucial question. Does the poorer prognosis after incomplete revascularization reflect more advanced disease, or is there a causal relationship between undergrafting and outcome? The multivariate analysis in our study indicated that undergrafting per se is an independent risk factor (hazard ratio, 1.82; 95% CI, 1.15–2.85) with a greater effect than smoking (hazard ratio, 1.57; 95% CI, 1.34–1.84) and diabetes (hazard ratio, 1.35; 95% CI, 1.18–1.54). This in turn provided indirect evidence that bypass surgery still does have a survival benefit. Otherwise, 1, 2, or more grafts would have no implication on outcome.
Off-pump Surgery
Off-pump coronary surgery has gained popularity because of an assumed reduced invasiveness by avoiding cannulation and extracorporeal perfusion. A number of studies have indeed shown that off-pump surgery is advantageous. Other studies have failed to detect a beneficial effect. Instead, many studies have reported a greater proportion of incomplete revascularization. Weather this has affected outcome has not been possible to establish. The effect of off-pump techniques on short-term and midterm outcomes has been reported in a great number of retrospective studies and in fewer prospective randomized studies. The results have been ambiguous.15-19 According to the present analysis, off-pump surgery does not by itself influence long-term survival because there was no significant difference in risk ratios of death between off-pump and on-pump techniques. The risk of a type II error has been considered. There is indeed room for a substantial difference within the CI (0.82–1.40), but the P value (P = .57, Figure 6) indicates that the risk of an error is not large.
However, off-pump surgery, according to our data, is associated with a higher risk of incomplete revascularization. There are 2 possible explanations for this. First, in some patients optimal surgical access is more difficult to obtain, and the surgeon might have refrained from grafting to avoid severe hemodynamic instability or the vessel might have been considered unsuitable because of poor quality. It is probable that these alternatives are less common at high-volume off-pump centers. It could even be argued that the low off-pump volume at our institution and relative inexperience makes it impossible to make any generalizations from our findings. This argument is contradicted by recent systematic reviews and meta-analyses of randomized controlled trials. These indicate that off-pump surgery is associated with more frequent incomplete revascularization, more frequent graft occlusion, and repeat revascularization than on-pump surgery.20-23
Second, in some patients with a higher risk profile, the off pump-technique was chosen primarily to minimize patient trauma by combining the off-pump technique with target revascularization. Then one of several diseased arteries, nearly always the LAD, was proclaimed the culprit lesion, and the other diseased vessels were left to minimize the procedure. This option complicates the interpretation of our data. However, because risk variables that were used to select patients for this strategy are included in the risk model, the role of incomplete revascularization as an independent risk factor can be determined. Caputo and coworkers,24 who also used an off-pump/target revascularization strategy, found that incomplete revascularization was associated with reduced survival, which is in agreement with our data.
Study Design and Limitations
The primary focus of this study was completeness of CABG versus long-term survival. For ethical reasons, a randomized design with complete versus incomplete revascularization is hardly permissible, and for this purpose, a registry observational study is the only possible study design. A secondary focus was the role of off-pump surgery. A prospective randomized study would have advantages in comparison with an observational study. A disadvantage, though, would be that a very large number of patients would be required. It would require more than 8000 patients to detect a difference in survival related to ungrafted vessels to obtain 80% power with a similar mix of patients as in this study. In reality, this means that the randomized approach will be impossible to pursue. The statistical methods of this study make it possible to identify independent risk factors against a variable background of a number of other risk factors to balance the disadvantage of the observational study. However, there were significant differences between the off-pump and on-pump groups in this study. This could have been a problem if there had been no overlap between groups. Fortunately, for our analyses, the overlap was extensive for all variables. The primary end point was death of all causes and was obtained from the population register administered by the Swedish National Tax Board, which registers all deaths occurring within and without all Swedish hospitals. Based on previous experience, complete follow-up can be assumed.
|
Conclusions |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. J. Head, M. J. Mack, D. R. Holmes Jr, F. W. Mohr, M.-C. Morice, P. W. Serruys, and A. P. Kappetein Incidence, predictors and outcomes of incomplete revascularization after percutaneous coronary intervention and coronary artery bypass grafting: a subgroup analysis of 3-year SYNTAX data Eur J Cardiothorac Surg, December 21, 2011; (2011) ezr105v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-T. Tsai, F.-Y. Lin, C.-H. Lai, Y.-C. Lin, C.-Y. Lin, and C.-S. Tsai On-pump beating-heart coronary artery bypass provides efficacious short- and long-term outcomes in hemodialysis patients Nephrol. Dial. Transplant., September 27, 2011; (2011) gfr536v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-H. Kim, D.-W. Park, J.-Y. Lee, W.-J. Kim, S.-C. Yun, J.-M. Ahn, H. G. Song, J.-H. Oh, J. S. Park, S.-J. Kang, et al. Impact of Angiographic Complete Revascularization After Drug-Eluting Stent Implantation or Coronary Artery Bypass Graft Surgery for Multivessel Coronary Artery Disease Circulation, May 31, 2011; 123(21): 2373 - 2381. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Ouzounian, A. Hassan, A. M. Yip, K. J. Buth, R. J. F. Baskett, I. S. Ali, and G. M. Hirsch The impact of sequential grafting on clinical outcomes following coronary artery bypass grafting Eur J Cardiothorac Surg, November 1, 2010; 38(5): 579 - 584. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Takagi, M. Matsui, and T. Umemoto Off-Pump Coronary Artery Bypass May Increase Late Mortality: A Meta-Analysis of Randomized Trials Ann. Thorac. Surg., June 1, 2010; 89(6): 1881 - 1888. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Hu, Z. Zheng, X. Yuan, W. Wang, Y. Song, H. Sun, and J. Xu Increasing Long-Term Major Vascular Events and Resource Consumption in Patients Receiving Off-Pump Coronary Artery Bypass: A Single-Center Prospective Observational Study Circulation, April 27, 2010; 121(16): 1800 - 1808. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. F. V. Panday, S. Fischer, A. Bauer, D. Metz, J. Schubel, N. E. Shouki, T. Eberle, and H. Hausmann Minimal extracorporeal circulation and off-pump compared to conventional cardiopulmonary bypass in coronary surgery Interact CardioVasc Thorac Surg, November 1, 2009; 9(5): 832 - 836. [Abstract] [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 |
