| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2007;83:1008-1015
© 2007 The Society of Thoracic Surgeons
a Centennial Heart Center, Vanderbilt University, Nashville, Tennessee
b Duke University Medical Center, Durham, North Carolina
c Drexel University, Philadelphia, Pennsylvania
Accepted for publication October 13, 2006.
* Address correspondence to Dr Rankin, 320 Lynnwood Blvd, Nashville, TN 37205 (Email: jsrankinmd{at}cs.com).
Presented at the Forty-second Annual Meeting of The Society of Thoracic Surgeons, Chicago, IL, Jan 30–Feb 1, 2006.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Methods: From 1984 to 1986, 867 patients with multivessel coronary disease underwent CABG. Single (SIMA) IMA grafts were used in 490 and multiple (MIMA) IMA grafts in 377, along with concomitant saphenous veins. Generally, MIMAs were placed to the two largest coronary systems. Among baseline characteristics, only smoking, diabetes, and hypertension were significantly higher for MIMA versus SIMA. Multivariable Cox model analysis was used to assess outcome differences between groups.
Results: During a median follow-up of 20 years, the composite of mortality, myocardial infarction, percutaneous coronary intervention, and redo CABG was significantly reduced after MIMA versus SIMA (p = 0.013). Event-free survival was extended by almost 1 year (p = 0.018), and redo CABG was reduced by 59% (p = 0.005). A comparison within the MIMA group was made between 235 patients receiving IMA grafts to left anterior descending/left circumflex territories versus 122 with grafts to left anterior descending/right coronary artery systems. No significant difference in composite outcome was observed between these configurations (p = 0.88).
Conclusions: These data confirm the clinical benefits of MIMA grafting in multivessel coronary disease to 20 years of follow-up. As long as MIMAs are placed to the two largest coronary systems, no significant differences in long-term results are evident between left anterior descending/left circumflex and left anterior descending/right coronary artery configurations.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Although the results of early studies were mixed, longer follow-up has clearly demonstrated significant benefits to MIMA grafting [2–4]. Our original study from 1984 was negative for many years [5, 6]; but with longer follow-up, it turned positive in the second decade [7]. At this point, 20-year data were available in our series, and the purpose of this study was to present outcomes of MIMA patients operated on at Duke University in the 1980s as well as to address questions about the optimal configuration of MIMA grafts. Specifically, the analysis was designed to quantify the magnitude of clinical benefits and to evaluate two different MIMA configurations.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
After a 3-year period, results with the MIMA approach seemed generally satisfactory, and a retrospective observational analysis was suggested to assess possible differences in early and late outcomes between SIMA and MIMA patients within the Duke Cardiovascular Databank. As near as could be assessed by retrospective review, the two surgeons exhibited statistically similar surgical philosophies, technical characteristics, and early outcomes. Detailed and specific comparison data are presented elsewhere [5]. From 1 to 32 weeks postoperatively, a 207-patient cohort from the MIMA group underwent coronary graft angiography under Institutional Review Board (IRB) approval and with informed consent. This represented an 85% restudy rate of patients operated on throughout that period. The angiography demonstrated a 98.5% IMA patency rate, suggesting good quality of MIMA grafts for testing of the hypothesis [9].
During the 3-year period, the first surgeon (emphasizing MIMA) performed 654 CABG procedures and the second (emphasizing SIMA), 413. Baseline clinical presentations, extent of coronary disease, and all other prognostically important characteristics were statistically similar in the two surgeons’ practices [5], but the first surgeon performed MIMA grafting in 71% of multivessel disease patients, whereas the second surgeon performed SIMA grafting in 74% of multivessel disease patients (p < 0.05). This finding again suggests a good test of the hypothesis. In addition, this approach tended to equalize baseline characteristics between groups and included a full spectrum of baseline patient risk in the MIMA cohort [7].
In the present analysis, patients from the two surgeons were pooled, and patients with single-vessel disease were eliminated because they were not candidates for MIMA grafting. Patients with multiple IMA anastomoses only to a single coronary system (as in sequential left IMA grafts to LAD and diagonal vessels) were considered SIMA patients, according to previous data demonstrating minimal prognostic effects of sequential grafts [7]. Standard pedicle grafts were performed predominantly, with only 15 free grafts in the study. The SIMA group comprised 490 patients and the MIMA group, 377. Within the MIMA cohort, 235 patients (62%) had IMAs to the left anterior descending (LAD)/left circumflex (LCX) systems, 122 (32%) had IMAs to LAD/right coronary artery (RCA) systems, and 20 (5%) had LCX/RCA IMA combinations.
During the 20-year follow-up, information was collected by the Duke Cardiovascular Databank on all-cause death, subsequent percutaneous coronary intervention (PCI), nonfatal myocardial infarction (MI), and redo CABG. Baseline demographics, clinical history, operative data, and immediate postoperative variables were assessed as percentages for discrete variables and compared using 
2 tests. The distributions of continuous variables were described as median (25th and 75th percentiles), and differences were assessed by Kruskal-Wallis tests. Cumulative event rates for each outcome, as a function of time after the index surgery, were calculated using the Kaplan-Meier method. Multivariable Cox proportional hazards modeling [10] was applied to a composite outcome [11] of all-cause death/PCI/nonfatal MI/and redo CABG in SIMA versus MIMA patients, adjusting for minor differences in baseline characteristics. The composite outcome was used to compare the 235 MIMA patients with LAD/LCX configurations with the 122 who had IMAs to the LAD/RCA systems. Because only 20 patients had the LCX/RCA configuration, these were not analyzed.
To perform adjustment procedures for each Cox model and for each outcome, the following input variables were included: age, gender, history of peripheral vascular disease, cerebrovascular disease, chronic obstructive lung disease, redo CABG, diabetes, prior smoking history, hypertension, mild or moderate valvular disease (left untreated), race; multivessel or left main coronary disease, acute presentation for the index surgery versus elective, New York Heart Association class III or IV congestive heart failure, ejection fraction, and SIMA versus MIMA grafting strategy.
All variables were dichotomized, with noncardiac comorbidities combined into a modified weighted Charlson index [12], and ejection fraction evaluated as more or less than 0.35. A restricted cubic spline function was used to define the point of change in the relationship between age and hazard of mortality, which proved to be 50 years. Thus, for ease of interpretation, the age variable was included in the model as older or younger than 50 years. A propensity score was generated for each patient, using all of the baseline characteristics and a logistic regression analysis, and the probability of receiving MIMA versus SIMA (from 0–1) was tested as a variable in the statistical model.
All covariates (each dichotomized, or with one degree of freedom) were included in the final Cox model, and statistical adjustment was performed even for small, insignificant differences (which in the aggregate could be important). A formal treatment interaction was performed between MIMA versus SIMA and the presence of diabetes, and also between MIMA versus SIMA and age younger than 50 years.
Unadjusted and adjusted Kaplan-Meier curves for freedom from the composite end point were generated, with the adjustment process accounting for minor differences in baseline characteristics. Hazard ratios for all variables as related to the 20-year composite outcome were tabulated, and the process was repeated for each component of the composite (expressed as cumulative incidence curves), and for the two configurations. Finally, differences in average life expectancy between MIMA and SIMA groups were assessed by comparing the areas under the individual adjusted survival curves. In each model, tests were considered significant at p < 0.05. This retrospective analysis was performed in the Duke Databank under a waiver from the Duke IRB.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
|
|
The reduction in the hazard for each adverse event ranged from 15% to 62% in the MIMA group, with composite outcome decreasing by 15%. Of special interest was the marked reduction in 20-year reoperation rates, from 12.2% in the SIMA group to 5.0% with MIMA.
|
|
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
In this series, it took 7 to 10 years of follow-up before an advantage of MIMA grafting could be appreciated [7], but from that point, and lasting to the 20-year mark, it is clear that MIMA benefits are statistically and clinically significant. This finding is similar to many recent studies, which viewed in the aggregate, virtually resolve this question [13–21].
All methods of clinical research have advantages and drawbacks, but the present study is somewhat unusual. Most centers that have examined this topic initially selected low-risk "ideal" patients for MIMA grafting and then compensated for the selection bias by observational analysis using Cox models or propensity score methodology. In this situation, questions exist about influences of confounding variables. With the present study design, however, the two groups had quite comparable baseline, operative, and early outcome variables. In the second decade of follow-up, however, benefits proved quite significant, with magnitudes of treatment effects that were similar to other reports [3], even though large statistical compensation maneuvers were unnecessary.
In fact, the MIMA patients in this study were slightly sicker than SIMA patients, a characteristic that could have tended to minimize therapeutic benefits. Thus, using a somewhat different analytical approach, but producing the same result, enhances the scientific strength of the MIMA grafting argument, given the principle that the more ways a subject is addressed the better. Finally, no differential effect was noted in diabetic patients and young patients in the treatment interaction analyses. The corollary of this finding is that higher-risk patients, such those with diabetes and elderly patients, benefited to a similar degree after MIMA grafting, suggesting that this approach is appropriate for the broad spectrum of coronary disease patients.
Most current studies suggest that MIMA grafting should be performed liberally in patients with multivessel coronary disease. Even in patients with diabetes, reduced ejection fraction, and emergent presentation, MIMA grafting can be accomplished with mortality and morbidity similar to SIMA. As with our early data [5], most recent reports do not identify an increased risk of sternal infection in diabetic patients, with or without the "skeletonizing" approach [22–26]. In some patients, small IMA size or specific coronary anatomy may limit IMA use, but more than 70% of multivessel patients in this series could undergo MIMAs, always remembering to assess pregrafting IMA function carefully to avoid the "bad IMA graft" syndrome [27].
At 15 years, 1 patient in 5 seemed to benefit from routine MIMA use [7], and average event-free survival over 20-years was extended by almost 1 year. These findings seem clinically important and again suggest that MIMA grafting should be applied broadly and considered the "therapeutic standard" at present. Some discussion has occurred about specific use of MIMAs for grafting left-sided vessels. In the present data, it appeared that LAD/RCA configurations performed just as well as long as the two largest vessels were grafted.
Questions exist at present about preferred revascularization strategies in patients with prognostically serious multivessel coronary disease [28]. In recent years, PCI with coronary stenting has been used more frequently, based on a perception that patencies with drug-eluting stents have improved. Recent data, however, indicate that long-term death rates with stents are higher than with CABG, especially in patients with three-vessel disease [29–32]. Moreover, published and early unpublished data on drug-eluting stents suggest that long-term results may be no better than with bare metal stents [33].
With the excellent late outcomes of the MIMA approach, it is suggested that MIMA grafting should be the primary form of revascularization for multivessel disease and that bilateral IMA grafts should be used in 50% to 70% of multivessel patients in current practice. More than half of patients in the present series eventually died of cardiac causes, however, which suggests that potential for future improvement still exists to diminish the phenomena of "second decade" cardiac complications. Thus, the current trend to expand arterial grafting further would seem appropriate by even more liberal use of IMA sequential grafts, free grafts when necessary, and IMA patch angioplasty. The addition of radial arteries as the primary adjunctive graft (with minimization of saphenous veins) in "all arterial" procedures may further enhance results [34]. After four decades of experience with CABG, it is striking that the techniques and benefits of surgical coronary revascularization continue to evolve and improve faster than ever.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
One of the things we have been interested in is the issue of "who benefits"? If the average patient has an improvement in outcomes are there nonaverage patients that fall on either ends of the spectrum who actually do worse with bilateral ITA grafting? In our study we identified elderly females as a group of patients who had poor outcomes with internal mammary artery grafting.
According to the STS database, less than 5% of patients currently receive multiple internal mammary artery grafts. Why is that? It may be that surgeons tend not to be rewarded for long-term outcomes. We are penalized for short-term adverse consequences but not rewarded for the long-term outcomes associated with the longer and more complex operation required to accomplish multiple internal mammary artery grafts. There also may be the thought that 20 years is such a long time that it is irrelevant to patients having bypass grafting. However, the patient that does not have a lot of comorbidity at the time of operation has a high likelihood of still being alive 20 years later, a point where the benefit of having multiple arterial grafts continues to be important.
My question for Dr. Rankin is, do you think that there are any patient subsets that did worse with bilateral ITA grafting, and do you think that there are patients with characteristics that you can identify that do particularly well with multiple arterial grafts, even better than the average?
DR RANKIN: I would like to thank Dr Lytle for his discussion, and recognize that these data are available because Dr Rob Califf and his co-workers have kept the Duke Databank going for over three decades now—a real accomplishment.
I just don’t know how we would look at elderly females within our Cox model analysis. Superficially looking at our data, I don’t believe we could define a group that did worse after multiple IMAs. Female gender doesn’t drop out as a significant factor, and advanced age doesn’t either. Perhaps we should do a treatment interaction with the combination of elderly and female. Certainly, the older people have less life expectancy, but the relative risk reduction with multiple IMAs versus single is the same as compared to the younger patients. So I guess at this point, I am back to performing multiple IMAs in 75% of multivessel patients, considering more the quality of the mammary arteries and the anatomy, than age or gender, per se.
It is a little more difficult operation, but it is well worth the effort in the long run. As you know, our data were negative for 7 years; but the reduction in cardiac events and death in the second decade of follow-up now seems irrefutable, as Dr Lytle has been saying for a while, and I believe the national incidence of multiple IMA grafting will increase dramatically in coming years.
DR ROBERT S. D. HIGGINS (Chicago, IL): Along those lines, your data showed that diabetic patients tend to have adequate lower infection rates and actually did well with multivessel revascularization. Can you comment whether you did anything particularly different with the diabetic patients who had multivessel revascularization?
DR RANKIN: Well, they didn’t do better. They had the same risk of sternal infection as the rest of the population. Certainly, they didn’t do worse. And I think that is supported now by multiple recent papers in the literature, Dr Endo’s paper, for example. In my practice, as evidenced by this series, diabetics have never experienced a higher sternal infection rate with multiple IMA grafting, and I don’t consider diabetes in the decision-making.
The issue of sternal infection is a whole other discussion, but if one pays good attention to all the technical factors that go into the prevention of sternal infection, it is possible to do multiple IMA grafts in diabetics without an increased risk of infection. In the past 10-years, my overall sternal infection rate has been less than 0.5%, and if it can be kept that low, MIMA grafting in diabetics does not seem to be an issue.
DR WILLIAM L. HOLMAN (Birmingham, AL): Scott, that presentation was a lot of fun to hear, having been under your tutelage during that era. I had two questions for you. First, were you able to break out the patients who had free IMAs as opposed to in situ grafts, and what effect did that have on patency, or was the number of free IMA grafts too small to really matter? Second, were there characteristics that distinguished patients who did not receive the multiple IMA grafts under the surgeon who was doing multiple grafts? Thank you so much for the paper.
DR RANKIN: Bill participated in a large number of these operations, so many that he maybe should be a coauthor, but thank you Bill for your contributions. As you know, we did a patency study that was published in 1986, and the free IMA grafts had excellent early patency, 100%; but there were only 15 free IMA grafts in that series, out of 338 IMA anastomoses angiogrammed. Because the free IMA is more technically difficult, with a proximal anastomosis and so on, we tended to emphasize the simplicity of pedicled grafts, then and now. Thus, the free IMA sample size is really not large enough to evaluate difference in long-term outcome.
The one thing we can say from the angiographic studies is that the right IMA to the LAD was virtually a 100% patency operation, as was the left in-situ mammary to the circumflex. So we tended to perform that operation when we did LAD-circumflex combinations, which accounted for two thirds of the multiple IMAs.
The one technique I liked in those patients was to use the thymic fat pad to cover the graft crossing the midline, and although the reoperation rate turned out to be very low, I think that protects the mammary from injury at reentry. In performing in-situ right IMA grafts to the distal RCA, we always did a "pulley stitch" of 3-0 pledgetted Prolene from the crux, to the lateral right atrium, and then to the aortic cardioplegia pledget, to take tension off the graft. This allowed the posterior descending to be revascularized whenever necessary, and probably contributed to the excellent RIMA patency we observed.
We haven’t specifically looked at the 29% of multivessel patients in my practice that did not get multiple IMAs, although they are included in the single IMA group. As you know, multiple IMAs usually were not done because the IMAs were small or for other technical reasons. But that is a great idea, and we may just go back and make sure those patients look the same as the rest of the group.
DR SUDHIR PREM SRIVASTAVA (Odessa, TX): Thanks for a great paper. I had a question. You have shown obviously very significantly improved survival with both mammaries and a reduced incidence of PCI or surgical interventions subsequently. Did you look at what grafts were occluded and what vessels needed intervention so maybe we can learn some lessons from that to even reduce that further in relation to multiple grafting with the mammaries?
DR RANKIN: Well, that also is a great idea, but we haven’t gotten into that. The data should be available, and we might explore that further. It is interesting here that a 20-year follow-up is really an examination of cardiac life history after coronary bypass, and there were some surprises: 70% of the patients were dead at 20 years, and 50% of those died of cardiac deaths. Now, we need to look at our definition of cardiac death in more detail and examine that variable closer, but I think analyzing the cardiac death aspect of it, the PCI aspect of it, and defining which vessels were required to have a late PCI, is a great suggestion.
DR THOMAS GLEASON (Chicago IL): Was there a difference in harvest technique between the two groups; ie, skeletonized versus non-skeletonized?
DR RANKIN: In this era, we did not do skeletonized IMA harvest; however, if there was difficulty reaching the distal vessel, we made liberal use of Dr. Cosgrove’s slit technique, and put multiple slits along the pedicle, which accomplishes pretty much the same thing. But in this early era, we did not use a skeletonized technique; all of these were pedicled harvests.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Tabata, J. D. Grab, Z. Khalpey, F. H. Edwards, S. M. O'Brien, L. H. Cohn, and R. M. Bolman III Prevalence and Variability of Internal Mammary Artery Graft Use in Contemporary Multivessel Coronary Artery Bypass Graft Surgery: Analysis of the Society of Thoracic Surgeons National Cardiac Database Circulation, September 15, 2009; 120(11): 935 - 940. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Tatoulis, B. F. Buxton, J. A. Fuller, M. Meswani, S. Theodore, N. Powar, and R. Wynne Long-term patency of 1108 radial arterial-coronary angiograms over 10 years. Ann. Thorac. Surg., July 1, 2009; 88(1): 23 - 29. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. A. Schwann, A. Zacharias, C. J. Riordan, S. J. Durham, A. S. Shah, and R. H. Habib Sequential radial artery grafts for multivessel coronary artery bypass graft surgery: 10-year survival and angiography results. Ann. Thorac. Surg., July 1, 2009; 88(1): 31 - 39. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Zacharias, T. A. Schwann, C. J. Riordan, S. J. Durham, A. S. Shah, and R. H. Habib Late Results of Conventional Versus All-Arterial Revascularization Based on Internal Thoracic and Radial Artery Grafting Ann. Thorac. Surg., January 1, 2009; 87(1): 19 - 26. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Bottio, V. Tarzia, G. Rizzoli, and G. Gerosa Total arterial revascularization, conventional coronary artery bypass surgery, and age cut-off for the loss of benefit from bilateral internal thoracic artery grafting Eur. J. Cardiothorac. Surg., January 1, 2009; 35(1): 191 - 191. [Full Text] [PDF]
|
|
|
|
|
|
D. Glineur, C. Hanet, A. Poncelet, W. D'hoore, J.-C. Funken, J. Rubay, J. Kefer, P. Astarci, V. Lacroix, R. Verhelst, et al. Comparison of Bilateral Internal Thoracic Artery Revascularization Using In Situ or Y Graft Configurations: A Prospective Randomized Clinical, Functional, and Angiographic Midterm Evaluation Circulation, September 30, 2008; 118(14_suppl_1): S216 - S221. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Daemen, N. Kukreja, P. W. Serruys, A. Abbate, M. J. Lipinski, K. J. Harjai, A. C.C. Ng, L. Kritharides, J. H. Burack, and E. L. Hannan Drug-Eluting Stents vs. Coronary-Artery Bypass Grafting N. Engl. J. Med., June 12, 2008; 358(24): 2641 - 2644. [Full Text] [PDF]
|
|
|
|
|
|
T. A. Schwann, A. Zacharias, C. J. Riordan, S. J. Durham, A. S. Shah, and R. H. Habib Does radial use as a second arterial conduit for coronary artery bypass grafting improve long-term outcomes in diabetics? Eur. J. Cardiothorac. Surg., May 1, 2008; 33(5): 914 - 923. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. H. Jones The Year in Cardiovascular Surgery J. Am. Coll. Cardiol., April 29, 2008; 51(17): 1707 - 1718. [Full Text] [PDF]
|
|
|
|
|
|
R. Martinez-Sanz, R. de la Llana, I. Nassar, and P. Garrido Use of Both Internal Thoracic Arteries in Diabetic Patients Ann. Thorac. Surg., February 1, 2008; 85(2): 690 - 690. [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 |
