CABG vs. PCI for Multivessel CAD
Advances in Stent Technology and Shared Decision-Making
Interview | In general, coronary-artery bypass grafting (CABG) has been shown with fewer revascularizations later on compared to percutaneous coronary intervention (PCI). However, with incremental advances in stent technology, is that still an accurate statement? CSWN: Interventions spoke with Sripal Bangalore, MD, MHA, director of research in the cardiac catheterization laboratory, director of cardiovascular outcomes group, and professor of medicine at the New York University School of Medicine, about his paper, “Everolimus-Eluting Stents or Bypass Surgery for Multivessel Coronary Disease,” published in the New England Journal of Medicine, about these recent developments in stenting.
CSWN Interventions: In the Everolimus-Eluting Stents or Bypass Surgery for Multivessel Coronary Disease (BEST) study, what were you looking at specifically?
Sripal Bangalore, MD, MHA: The basic reason to do this study is if you look at prior randomized trials, or at least in a couple of trials, there is a signal that bypass surgery is associated with better mortality when compared to PCI. We saw those findings in the FREEDOM trial and the longer-term results of the SYNTAX trial, but it was actually at a mortality benefit. But all of the trials compared bypass with first-generation stents, and we know from plenty of data that the second-generation stents are far superior to first-generation. So we wanted to see the outcomes of comparing second-generation stents, and that’s the reason we did the study.
Now, in this particular analysis, what did you learn?
A couple of things. We had over 18,000 patients—a large number—and this large number of patients is especially important because we can actually look at individual endpoints. Historically, what clinical trials have done is to use composite endpoints. Composite endpoints are great for clinical trials because they make the sample size smaller, and we can do the trials easily. But in clinical practice, we don’t use composite endpoints.
We don’t tell our patients, “By the way, your risk of events is reduced.” They don’t understand what it means. In clinical practice, we look at individual endpoints. We tell the patient, “Your risk of death is this. Your risk of MI is this.” But clinical trials have never been powered, so here we had a unique opportunity to look at 18,000 patients and ask, “What are the individual endpoint difference between bypass and the latest generation stents?”
The results are very interesting, and it is actually very similar to a small randomized trial, which was presented at ACC. What we found was, for death, there was no differences between bypass surgery and stenting, and the same was seen in the BEST trial. So no difference—that itself is very interesting because of the signal we have seen in prior randomized trials favoring bypass. For myocardial infarction (MI), we showed that there was an increase in MI with using stenting compared to coronary-artery bypass grafting (CABG), but this increase in MI was not seen if patient had complete revascularization in the PCI group. So there is an increase in MI. But if you were to completely revascularize, no more differences in MI.
The BEST trial also showed increase in MI with the everolimus-eluting stent, but researchers didn’t do this additional analysis of looking at completeness of revascularization. The third thing was repeat revascularization. Both the studies showed that there is increase in repeat revascularization with PCI, but we did an interesting analysis. There have also been multiple publications from the New York State registry, and many of them have been published in the New England Journal of Medicine in the last couple of decades. We show a considerable improvement in stent technology in the rate of repeat revascularization from the balloon angioplasty era, bare-metal stent, drug-eluting stent, and with the latest generation stent. We show that there is an incremental improvement. And, finally, for stroke, we showed that there is increase in risk of stroke with bypass, which has been show in every other study.
For whatever reason, the BEST trial did not show a difference, but the BEST trial is very underpowered to actually look at any kind of event. We also looked at short-term outcomes, and we found that bypass surgery was associated with increase in the risk of death and stroke at 30 days, which has also been shown in prior studies.
In an accompanying commentary, Robert Harrington, MD, FACC, emphasized that there are still observational studies, and so that needs to be considered. With all the sophisticated statistics that now can go along with this, are we getting closer to the situation in which it is not a randomized trial, but one in which it is superior to what we used to know as just an observational study?
Both randomized trial and observation studies have their pluses and minuses. As you know, randomized trials are the gold standard. I don’t think we can ever replace that.
But also we have to be very aware that there are good randomized trials, there are bad randomized trials, and if the sample size attained is considerably smaller than what it was powered for, you are going to have a very underpowered study. The randomized trial is becoming as problematic as any other data out there.
Observation studies have advantages, and with using sophisticated techniques we can adjust for baseline confounders, just as we did, but we can be powered for individual outcomes. My take-home for all of this is that we need to consider the totality of evidence. We need to look at observation studies. We look at randomized trials, and based on everything we need to make a decision as to what does the data show us.
That’s exactly Dr. Harrington pointed out: there are clearly trade-offs between these two revascularization strategies that need to be discussed with the patients and part of the shared decision-making that is recommended by the guidelines.
I think that’s the critical thing to emphasize. If there is a mortality difference—mortality trumps everything—everybody should be going for bypass. So now we are saying, both from our study and the BEST study, there is no mortality difference. So mortality advantage of bypass is off the table. Now, it should be a discussion between patients and physicians that there are tradeoffs. With bypass surgery there is upfront risk of death and stroke; with the PCI there is long-term risk of repeat revascularization and possibly increased risk of MI if you cannot completely revascularize. So patients have different opinions. There are some patients who do not want to have a stroke regardless of how small the risk is, and there are some patients who do not want to come back for repeat procedures. So I think it ultimately becomes a shared-decision making.
Bangalore S, Guo Y, Samadashvili A, et al. N Engl J Med. 2015; 372:1213-22.
Harrington RA. N Engl J Med. 2015; 372:1261-3.
Editors’ Note: This interview has been edited from transcript.
< Back to Listings