Author + information
- Robert A. Guyton, MD∗ ()
- Carlyle Fraser Heart Center, Division of Cardiothoracic Surgery, Joseph B. Whitehead Department of Surgery, Emory University School of Medicine, Atlanta, Georgia
- ↵∗Reprint requests and correspondence:
Dr. Robert A. Guyton, The Emory Clinic, 1365 Clifton Road, Atlanta, Georgia 30322.
- coronary artery bypass
- coronary revascularization
- drug-eluting stents
- percutaneous coronary intervention
There is increasingly compelling evidence of the advantages of multiple arterial bypass grafting compared with the use of a single arterial conduit (1–4). In this issue of the Journal, Habib et al. (5) have made another contribution to the weight of this evidence. There are, however, concerning limitations to this single-institution retrospective study.
As we evaluate the weight that should be given to this study, we must ask 2 critical questions. First, what were the patient-specific pathologic factors that determined the choice of revascularization procedure? Were these patient-specific factors appropriately measured and recorded to allow reasonable risk adjustment of the 4 groups compared? Second, is the single institution sufficiently representative that the conclusions of this study are applicable elsewhere?
The study was conducted at the Beth Israel Medical Center in New York. This is an institution with a long tradition of excellent coronary revascularization, with superb results over the past quarter century (6,7). Surgeons at this institution have been champions of the radial artery as a second arterial graft. As percutaneous coronary intervention (PCI) emerged, the superior PCI practitioners at the Mount Sinai family of hospitals aggressively pursued off-label stenting for multivessel coronary disease. Stenting became the default therapy for multivessel disease.
In the patients eligible for this retrospective study, we must be honest about the factors that determined the choice of revascularization procedures. Especially after drug-eluting stent availability, multivessel stenting was performed in most patients in whom a PCI could accomplish substantial revascularization. This fact is confirmed by the data that the authors present. In the 11 years from which the PCI data were derived, 417 multivessel study–eligible patients on average per year underwent PCI. In the 17 years of coronary bypass grafting (CABG) data collection, 224 patients on average per year underwent CABG and approximately 30% of these had left main artery disease. This means that in this institution more than twice as many multivessel, non–left main artery patients underwent multivessel PCI as CABG. This very aggressive use of multivessel stenting was dramatically higher than the use of multivessel stenting in the State of New York reported for bare metal stenting (1997 to 2000) and for drug-eluting stents (2003 to 2004) (8,9).
Review of the study’s demographic data reveals that the choice of patients for stenting versus CABG was determined primarily by the extent of coronary disease. The real difficulty with the data collected is that the specific coronary pathology that determined the use of stenting, single-artery grafting or multiple-artery grafting, was neither measured nor recorded in the data available for statisticians to make appropriate adjustments. The data collected did not distinguish between easily stentable coronary pathology, arteries appropriate for a second arterial graft, and arteries not appropriate for a second arterial graft. These granular data on coronary pathology are most important. The triple-vessel cohort of SYNTAX (Synergy between PCI with Taxus and Cardiac Surgery) demonstrated that progressively advanced coronary disease treated by either PCI or CABG is associated with progressively increasing intermediate and late mortality (10).
We are left then with 3 important facts. First, the most discriminating factor in the choice of revascularization procedure was not recorded. Second, this important factor is strongly associated with intermediate and long-term mortality. Third, as the data collected (double-vessel vs. triple-vessel vs. left main artery disease) only minimally reflect this important factor, no amount of statistical gymnastics can reliably adjust for this missing variable.
Habib and colleagues got it right as they describe the limitations of their study: “Our study findings are based on a retrospective analysis of PCI and CABG data from a single institution, which may limit their generalizability. Propensity matching may not have accounted for unmeasured confounders.” They subsequently make a quite zealous conclusion, implying that their data are the source of this conclusion. Their analysis is substantially confounded by a strong institutional preference for multivessel stenting, by the absence of sufficient granular data on the extent of atherosclerotic coronary artery disease, and by the institutional policy of steering patients younger than 65 years of age with specific anatomy suitable for radial artery grafting to multiple arterial coronary bypass grafting. I believe that the conclusion is correct, but it is justified by the weight of evidence from multiple studies, not the data and analysis presented.
The authors pointed out that their findings are limited in generalizability. I would argue that the generalizability of techniques in coronary artery revascularization is both a problem and an opportunity. When we are dealing with medical procedures that are often highly technical with outcomes dependent on both operator skill and technique-specific experience, there should not be a huge attempt to generalize the procedure pathway. We seek to provide the best procedure for the specific patient, for the specific pathophysiology, and in the specific circumstance in which the decision must be made. This circumstance includes knowledge of local outcomes with various procedures. Shared decision making in the case of medical interventions that are experience and skill dependent includes accommodation of the expertise and preferences of the operators who are available to perform those interventions.
Consider an illustrative example: the ERACI-II trial reported a 5.7% 30-day mortality rate for CABG and a 0.9% mortality rate for PCI (11). These outcomes are very different from the equivalent CABG/PCI mortality in the current study. If I had multivessel disease and found myself in one of the ERACI-II hospitals at that time, please allow me to choose PCI. If I am in the Beth Israel Medical Center and my pathology is suitable for either PCI or single or multiple arterial coronary bypass, please let me have multiple arterial coronary bypass. Shared decision making includes an understanding of the skill and the expertise and the outcomes that are reasonably available to the patient if an intervention is being considered. We know this. We rarely verbalize it.
In summary, these authors have added to the weight of evidence favoring multiple arterial bypass grafting. They appropriately state limitations. As these limitations are considerable, the conclusions are overly zealous and might best be restated as follows: “Our experience is consistent with other studies indicating that survival for revascularization from coronary artery bypass including at least 1 arterial graft is superior to survival after multivessel stenting except in the earliest stages of multivessel disease. Our data indicate that a second arterial graft in patients younger than 65 years of age confers a particular survival advantage compared with multivessel stenting when using bare-metal or drug-eluting stents. At our institution, we have found the radial artery to be an effective second arterial conduit when applied to a second arterial target with at least a 70% proximal stenosis.” This conclusion more accurately reflects the state of the art.
Take heart, readers! There is a prospective, randomized study that will shortly help illuminate this issue. Professor David Taggart’s ART study of single versus multiple arterial coronary grafting has completed enrollment and is eagerly anticipated (12).
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Guyton is a member of the Valve Advisory Board and serves as a consultant for Medtronic, Inc. Richard Shemin, MD, served as Guest Editor for this paper.
- American College of Cardiology Foundation
- Locker C.,
- Schaff H.V.,
- Dearani J.A.,
- et al.
- Dorman M.J.,
- Kurlansky P.A.,
- Traad E.A.,
- et al.
- Habib R.H.,
- Dimitrova K.R.,
- Badour S.A.,
- et al.
- Mohr F.W.,
- Morice M.C.,
- Kappetein A.P.,
- et al.
- Rodriguez A.,
- Bernardi V.,
- Navia J.,
- et al.
- Taggart D.P.,
- Altman D.G.,
- Gray A.M.,
- et al.