Conversations in Cardiology: Q&A
Dr. Kern,
Just wanted your opinion about an ongoing debate at our hospital regarding CAD, in particular proximal LAD disease +/- DM. There is an ongoing argument between CV surgery and IC (Interventional cardiology, mostly me) re: treatment. The surgeons argue that all proximal LAD (isolated or multivessel), particularly in diabetics, should go for bypass; and they have "read the literature."
I have argued that:
a) not all disease is the same, particularly not all 3 vessel disease, and that we should be using a scoring system (such as SYNTAX) if we want to be academic
b) guidelines are not definitive: CABG does have advantage of decreased TLR driving MACE, but is there really long term survival benefit (with exception of BARI)? There seems to be catch-up with bypass as there is late SVG disease.
c) FREEDOM trial: seems to be driven by TLR and non-cardiac events
d) Other trials such as ARTS 1, ARTS 2, etc. seem to be about the same (except for TLR)
Curious as to your thoughts and debate at your institute. Might be topic for Conversations in Cardiology (which by the way is a great segment you print)
We have tried the "Heart Team" approach, and it is very one sided.
Thanks
Zaheed Tai
Bostick Heart Center
The issues for comment as you present them appear as
1. All proximal LAD (isolated or multivessel), particularly in diabetics, should go for CABG (according to surgeons).
2. All proximal LAD is not the same, and we should use SYNTAX to make better decisions (according to interventionalists).
3. If one excludes TLR, survival is similar as noted in FREEDOM, ARTS 1, ARTS 2, etc (not BARI); and therefore, PCI should be equally considered.
4. Is the Heart Team approach 'one sided' on this issue?
My view: CABG should be reserved for intermediate to high SYNTAX score in most people (considering additional clinical needs), but more so in diabetics. A low SYNTAX score with proximal LAD disease may strongly benefit the patient since it saves CABG for later in life when the disease progresses. Many of the SYNTAX scores could be moved to lesser risk if FFR is used where appropriate.
As for the Heart Team approach, this depends on the quality of the surgeon and the honesty of both the interventionalists and surgeons to understand the outcomes in the same light. If I asked the surgeon at your place, would he have made the same comment?
Let's see what the others say. Thanks again.
Mort Kern
This is an issue that we have struggled with as a subspecialty since the beginning. It is difficult to compare the relative risks and benefits of PCI versus surgery in an unbiased manner. If there is no survival difference, only TLR, most patients would rather have an angioplasty, even several repeat angioplasties, than have open heart surgery. Even if there is a survival advantage, depending on how much it is, most patients would still rather have a PCI. So I discuss these issues with the patients and ask them to speak with our surgeon as well to make an informed decision. Most still choose PCI. For those diabetics with diffuse disease, long CTOs, LM bifurcations, I recommend surgery. If the patient is still reluctant to have CABG, then I tell them I am willing to perform angioplasty as long as they know that the risk of recurrent PCI is higher. In addition to an interventionalist and a surgeon speaking to the patient as a "Heart Team," it might be better to have a cardiologist who is not an interventionalist on the "Heart Team" who can speak with the patient with less bias.
Jonathan Tobis
Agree with Jon…
Of course, what you are describing, Jon, is "shared decision-making," which should be a cornerstone of all that we do. We should embrace that practice for all of the therapies that we offer as providers. Much has been written about Shared Decision Making (SDM)…there is even a non-for-profit foundation with that name here in Boston. SDM implies a patient-centric approach, imagining all that you would want to know as a patient in order to make the decision with your physician or other provider. SDM also requires that the patient is "informed." That, of course, is where things can get a bit dicey, or at least highly variable. Even without placing value on the evidence (which we all inevitably do), there is a high degree of variability in interpretation of the data. Part of this is because the data are incomplete and are based on large trials, and they are not always generalizable to the types of patients who were NOT enrolled in the study (selection bias may distort trial results and interpretation). Nor do these large clinical trials always "speak to" the individual patient sitting in front of us. So, the "informing" of patients can be highly variable. That can be seen within our specialty (e.g. those who are passionate believers in COURAGE will deliver the information in different ways than those who believe the study was fundamentally flawed and not generalizable). The differences in interpretation and "information provided" can be accentuated even more when coming from different specialties, given that the self-interests of the specialties may diverge. The Heart Team approach is intended to get interests more aligned and in sync with those of the patient. However, in order for this to work, the incentives to work together and come to consensus about information and interpretation of outcomes… and what is best for the patient… must also be aligned. I personally believe for Heart Teams to ultimately succeed (and same is true for any other conjoined team approach amongst physicians), there must be some form of integration - economic for sure, but even more than that. Members of the team need to be equally invested in the process and believe in what they are doing and that it is ultimately for the good of our patients, for our practices, and for us as individuals as well. Sorry to wax on about this, but Dr. Tai has highlighted an issue that truly deserves our attention. We need to work hard to come up with models that will enable us to practice in the most positive, ethical, and non-fractionated way. After all, this is what we all would want if we were patients ourselves, and our patients should expect no less. As to the issue of the data for isolated proximal LAD disease, I believe the results are similar enough that the decisions should be individualized for each patient, based on a shared decision-making model, as Dr. Tobis describes, with balanced discussion about merits and liabilities of each approach, both of which are reasonable.
Ken Rosenfield
Well said, Ken.
Larry S. Dean
University of Washington School of Medicine
These important topics that have engendered debate in several venues.
- For the topic of revascularization for isolated disease of the proximal LAD, all evidence from randomized trials (Refs. 1-6), meta-analyses (Refs. 7-9), and a forest plot (Figure 1) suggests equivalence between PCI and CABG.
- For the topic of revascularization in diabetics with multivessel CAD involving the proximal LAD, an updated guideline has a dedicated section (Ref. 10). Please note that in the FREEDOM trial, the mortality difference between CABG and PCI at 5 years was associated with a P value of 0.049 that, because of a statistical penalty for interim analyses, would not have met the prospectively defined level of significance for the primary end point (P <0.044).
1. Thiele H, Oettel S, Jacobs S et al. Comparison of bare-metal stenting with minimally invasive bypass surgery for stenosis of the left anterior descending coronary artery: a 5-year follow-up. Circulation 2005;112:3445-50.
9. Kapoor JR, Gienger AL, Ardehali R et al. Isolated disease of the proximal left anterior descending artery comparing the effectiveness of percutaneous coronary interventions and coronary artery bypass surgery. JACC Cardiovasc Interv 2008;1:483-91.
10. Fihn SD, Blankenship JC, Alexander KP et al. 2014 ACCF/AHA/AATS/PCNA/SCAI/STS focused update of the guidelins for the diagnosis and management of patients with stable ischemic heart disease. Circulation 2014: in preparation.
11. Farkhouh ME, Domanski M, Sleeper LA et al. Strategies for multivessel revascularization in patients with diabetes. N Engl J Med 2012;367:2375-2384.
John Bittl
What I find most interesting about the question posed is that although the Heart Team approach would seem to be the most patient-centered manner to make these decisions, in fact, it allows any consultant with intransigent opinions who is not cognizant of the literature to make assertions that are not evidence-based. Decision by committee naturally leads to politics and interpersonal interactions and not necessarily more rational decisions. One problem is that surgeons aren’t typically in the position of making these decisions routinely; and naturally, they find surgery optimal when it is low risk. If you ask a barber if you need a haircut, you always do. If you ask a surgeon if you need an operation...
Interventionists have biases, too; but we are trained to weigh the relevant facts, and we struggle with the evidence. Most of us do our best to advise the patient appropriately. One of the real problems is that surgeons look only at mortality endpoints and don’t consider stroke, MI, quality of life, angina relief, atrial fibrillation, pulmonary complications, or pericarditis as real issues. I think we are going to see more of this going forward.
Lloyd W Klein
Rush Medical College
Thanks to John Bittl for a concise summary of the literature and current state of the art. No surprise that John nails it.
Lloyd, we all have our biases (except ME, of course!)
So, the challenge for us is how to standardize the approach to patients in a way that minimizes bias and maximizes support for the patient and helps them make tough decisions. When the data are clear, it is pretty obvious and easy. Definitely more challenging when the data are not available or clear. It is not that decisions need to be made by committee nor that we always need to be in agreement, but we do all have to work together; and the better we understand others' points of view, the higher road we take.
Ken Rosenfield
Well, the surgeons you guys work with must be different from those I encounter. I respect greatly their opinions as to technical feasibility and operative risk. Those opinions are always useful and usually accurate. For the most part, they realize that choosing among the three strategies available in 3 vessel disease requires familiarity with their pros and cons and that they are expert in just one. The posed question shows how things function in real life when surgeons are placed in a position of strategic choice that has conflict of interest engraved on it; and I stand by my apprehension that a surgeon who thinks the literature shows LAD lesions are best treated with CABG represents just the tip of this iceberg. While Kenny and Larry hope that a Heart Team will balance out biases and lead to a standardized decision of high scientific merit, I cannot accept that such choices, in all but the most clear cases, even exist in multivessel CAD. The optimal application of SYNTAX, FREEDOM, ASCERT, and COURAGE isn't at all defined in most cases and that leaves out FAME-1 and FAME-2. I believe patients’ interests are best served by addressing the pros and cons of all 3 choices and assisting them to find the best choice for them rather than a committee making that choice for them. I like my patients to meet the surgeon even if PCI is preferred. But I think a quixotic notion of "best choice" is actually dependent on many circumstances, and a highly individual one in most complex cases. I suppose I think cardiology is more art than science and recognize that this places me in the minority; but nevertheless to me, that's what makes our field so challenging and interesting. Our disagreement rests on what working together means and the most effective process to reach a decision with the patient foremost. Our guidelines, AUC (appropriate use criteria), and textbooks provide an evidence base we can rely on. But, patient preference and their assessments of risk and benefit, and their risk averse attitudes, also have a critical influence. Presumably, they are "on" the committee too.
Lloyd W. Klein
This is a very important discussion. It is striking that there are so many situations where clinical equipoise is still present, despite SYNTAX, FREEDOM, COURAGE, and other trials. All of these studies also have unknown relevance to today's technologies as they were performed with first generation DES (or BMS), and weren't necessarily ischemia guided. CABG and medical therapies also continue to improve. We are trying to reduce the equipoise by performing EXCEL and ISCHEMIA, both critically important trials that will answer many of the questions we are all struggling with. It took years to raise the hundreds of millions of dollars (!) to be able to perform these trials to address these issues. Yet we are having trouble enrolling patients in each. Research in the US is dramatically declining; and with the clinical equipoise that exists on these topics, I would have thought that the community would embrace the opportunity to generate the data we desperately need. I would like to appeal to everyone on this thread who are participating in one or both of these trials to personally become involved at your institutions and make it a priority to help us complete enrollment in these studies so we can develop this critical contemporary evidence to guide optimal patient care decisions.
Gregg W Stone
I couldn't agree more with Gregg. These are critical questions that we need answers to. The current trials (typically enrolling low in US) are critical to get us closer to an answer.
Roxana Mehran
I would recommend, and I believe the data support, PCI as preferable in patients with SYNTAX scores of 22 or less irrespective of diabetic status and CABG as preferable for patients with SYNTAX scores of 33 or greater irrespective of diabetic status. For patients with mid-tercile scores of 23-32, PCI may be preferable in the absence of diabetes and may be a reasonable alternative strategy in the presence of diabetes for patients at high surgical risk, if the specific center experience/expertise with PCI is favorable. It is noteworthy that in the 3 vessel disease cohort (non-LM), there was no difference in the composite of death/MI/stroke among the lowest tercile score patients through 5 years (p=0.56). These recommendations take into account the fact that both SYNTAX and FREEDOM may be considered obsolete due to the use of first generation DES as well as the lack of use of either FFR (ischemia guidance) or IVUS optimization. EES have demonstrated significant reductions in MI (vs. non-EES DES), ischemia driven-TLR (vs. PES), and stent thrombosis (vs. non-EES DES). This has been demonstrated in individual randomized clinical trials and pooled analyses of randomized clinical trials as well as network meta-analyses. Both FFR and IVUS have demonstrated significant reductions in death/MI through 2 years follow up in DES treated patients when compared with angiography and no IVUS, respectively. Finally, the specific level of center expertise is critically important. Remarkably, the intercenter variability in MACCE through 5 years in SYNTAX for a specific modality (CABG or PCI) exceeds any observed difference between CABG and PCI for the trial as a whole. Some centers do extremely well with PCI and others extremely well with CABG, and the outcomes are not correlated with trial enrollment volume. I would gladly have my admininistrative asstistant send all references in support of these comments tomorrow morning. I am painfully typing this myself. EXCEL will hopefully give us better insights as to the results with optimized PCI.
Dean Kereiakes
Please note that in the excellent SYNTAX Score II manuscript (Lancet 2013), diabetes was NOT a predictive factor favoring CABG over TAXUS. Rather, it was all about anatomical complexity and other factors which may be worse in diabetics. FREEDOM was driven by the fact that 85% of patients had 3 vessel disease, with a higher mean SYNTAX score than in SYNTAX! Patient with diabetes and anatomy favorable for PCI have excellent outcomes, with results likely even better with current generation DES.
Gregg W Stone
Also note that the AUC (Appropriate Use Criteria, Figure 2), although in need of an update, don’t find diabetes or LVEF to be an important variable in decision-making. Most of the decision seems to be based on anatomical complexity.
Eric Bates
I continue to marvel at just how wide the gaps are between guidelines, sub-specialty opinions, data, and real world practice continue to be.
This has been a long and exceptional discussion based on the data available on anatomic selection, SYNTAX score, etc. In the real world, however, the bedside question is a bit more than simply the academic perspective.
In the real world there are two additional factors that continue to stand outside of both available data sets and guidelines:
1. Application of anatomic considerations in patients with multiple co-morbidities (not just diabetes): CABG vs. PCI studies since BARI have all suffered even in the planning phase on identifying the patient profiles acceptable for both therapies not only based on coronary anatomic subgroups but on non-coronary co-morbidity profiles… renal failure, COPD with very low FEV1, bleeding diatheses, stroke history, etc. In real practice these patients constitute a large number of scenarios where the anatomy, based on randomized clinical trials (RCTs), might favor CABG; but good luck finding a surgeon to do the procedure.
2. In the era of the internet, patients have their own strong and well-formed opinions; and while informed by data, we can apply probabilistic conclusions to our patient conversations. And no RCT has really addressed the common patient inquiry: “what if we try PCI first, using the best of new DES platforms, and if there is restenosis (noting that repeat revascularization is the primary driver of most PCI vs CABG statistical study differences from POBA to BMS to SYNTAX and FREEDOM) then revisit this question….” This is frequently the "look you in the eye" patient key question: “Doc, if this was your anatomy, what would you have done?”
I think this identifies an real arena for "translational" research, e.g.: what is the real tool kit for bridging the bedside moments where academic data, practice guidelines, and the application to a single human being do not well line up the sun, the moon, and the stars...
My 2 cents.
Mitchell Krucoff
Comments