Two years ago, the first Grand Debate was held, and the organizers are excited to call it a tradition and continue it. Chairpersons Eugene Braunwald (Boston, US) and Frank Ruschitzka (Zurich, CH) expressed their hopes for a true battle between opinions, as a means to learn more about the field. By taking extreme positions, the debaters showed different ways that the data can be looked at, and sometimes misinterpreted. Debates like this can help look at the data critically.

To position the topic, Eugene Braunwald went back in time, and remembered a publication of 1982 in which he and Robert Kloner wrote about the ‘stunned myocardium’, referring to disturbed structural, metabolic and functional properties of reversibly injured myocardium that may last for days, which occurs after brief periods of non-lethal ischemia. They also wrote that when the myocardium is repeatedly stunned, it may exhibit chronic post-ischemic left ventricular (LV) dysfunction, which can progress to ischemic cardiomyopathy. The question arose whether chronic LV dysfunction could be ameliorated by improvement of myocardial perfusion. Coincidently or not, that is also the objective of this discussion. Many people at the time studied myocardial viability. Indeed, a study (Allman, JACC 2002) showed that if the myocardium is still viable, revascularization improved survival, while if it was nonviable, the intervention did not affect survival. A later study by Bonow (NEJM 2011) also demonstrated an association between viable myocardium and better survival, but after adjustment for other baseline variables, the observed effect was no longer significant. Thus, concluded Braunwald the introduction, we still have a lot to learn.

John Cleland (Glasgow, UK) defended the PRO-position on the statement ‘Coronary revascularization provides no meaningful clinical benefit over optimal medical therapy (OMT) in patients with stable ischemic HF. In order to convince the audience that there is no benefit of this intervention, he first showed that in about half of patients with ischemic heart disease with a previous myocardial infarction (MI), MI is silent. Thus, it is important to weigh risks and benefits of revascularization. Potential benefits include relieving angina, improving HF symptoms and LV function, preventing ACS/angina and increasing longevity. Revascularization can, however, also lead to procedural deaths and myocardial damage, or the procedure may be futile. Revascularization has been described as a ‘controlled’ MI, with substantial infarct sizes and myocardial scars as a consequence of revascularization.

Does it relieve angina though? Statistically significant improvements have been shown with, as compared to without CABG, but is this clinically meaningful? About that, Cleland is not so sure.

Unfortunately, lack of ischemia, nor myocardial viability also did not prove to be indicators of greater benefit of revascularization. And symptoms, are they relieved by revascularization? Some improvement has been described, but this was in an unblinded study (Stewart, Open Heart. 2018). Moreover, the difference between CABG plus OMT and OMT was modest. Not much effect on the 6-minute walking test was seen, although a Hispanic, Latino and non-white subgroup seemed to experience a bigger effect.

With regard to functional effects, observational studies have shown improvement of left ventricular function, but at percentages that can also be obtained with drugs. And, measurement errors cannot be ignored in that setting. In a small RCT, if anything can be concluded from it, medical therapy appeared better than revascularization (ESC Heart Failure, 2014), possibly due to the damage caused by the procedure

The STICH trials suggested that patients on medical therapy experienced more events, but what the event curves do not show is that 100 patients randomized to OMT crossed over to CABG, which could have accounted for the observed differences. The STICHES trial showed that maybe in the long run, a modest benefit of CABG may be expected over OMT, but the HEART and STICH trials did not show a benefit in shorter trial duration. Interestingly, in subgroup analyses of STICH, patients in ‘other countries’, benefitted more. These were Indian patients; are surgeons in India particularly skilled?

In all of these studies, patients were not on ARNI treatment yes; this may also have a similar effect. Moreover, most deaths in the STICH trial were sudden death, so possibly more ICD’s would have prevented those deaths. STICHES did show a possible mortality benefit in those under 55 years, and again most of the benefit was seen in patients in India.

Thus, Cleland concluded that revascularization can reduce mortality somewhat, provided that age is <55 years, patients have three-vessel disease, very poor LV function, no device therapy, and that patients are willing to take the pain and risk of CABG. Little effect can be expected on symptoms, and there is no effect on functional capacity. Only anecdotal evidence on effects on LV function is available. Thus, as always, more research is needed (REVIVED-BCIS2 is ongoing, focussing on PCI).

Eric Velazquez was invited to defend the CONTRA position. He started out by showing a graph of the O’Connor study (Am J Cardiol 2002), in which a lot of the early work of the Duke CVD Databank was visible, namely that a long-term survival benefit of CABG over medical therapy was seen in patients with ischemic cardiomyopathy, and this benefit was rapidly visible. It seemed particularly true for patients with low ejection fraction, but this remained to be proven.

The -unblinded- STICH trial randomized patients to medical therapy or guideline-directed medical therapy plus CABG. In the first 5 years of follow-up, a non-significantly reduced all-cause mortality risk was seen (HR: 0.86, 95%CI: 0.72-1.04, P-log-rank: 0.123). When follow-up was extended with 5 years, a similar, but statistically significant, risk reduction was seen (HR: 0.84, 95%CI: 0.73-0.97, P-log-rank: 0.019). Velazquez did not know about another HF trial that showed an absolute risk reduction of 8%, which is clinically and statistically significant. Significantly more people who also underwent CABG showed a significant change in quality of life, up to 36 months.

But, it should be noted that these results, all in favor of CABG, were intention-to-treat analyses. 55 Patients randomized to CABG never underwent the intervention. And 65 randomized to OMT only, received CABG. In per protocol analyses, the effect was smaller. In the spirit of the debate, Velazquez pointed out that dr. Cleland co-authored an article in which it was indeed concluded that the cross-over events diminished the impact of CABG on survival as seen in the intention-to-treat analyses. In other analyses of STICH, those without angina at baseline showed a larger mortality benefit with CABG, although when cross-overs were considered, the risk reduction was of similar magnitude in those with or without angina at baseline. Overall, based on STICH it was concluded that CABG leads to earlier and greater benefit in those at highest risk, namely those with 3-vessel disease, EF <27% or LVSVI >79ml/m2.

A meta-analysis of 21 studies including 16191 patients comparing CABG, PCI and OMT (Wolff, Circ HF 2017) confirmed the benefit of invasive revascularization strategies, with HRs for CABG vs OMT of 0.66, PCI vs. MED of 0.73 and CABG vs. PCI of 0.82.

It should also be noted that ischemic heart disease in association with HF impacts the most socioeconomically deprived subset of the HF population. These are not the first patients who are tested for CAD and subsequently receive optimal treatment. Thus, there is a large opportunity for improvement. In another article (Ferreira et al.), Velazquez pointed out, Cleland wrote that performing coronary angiography was associated with improved outcomes in worsening HF. Thus, addressing coronary artery disease as a therapeutic target in worsening HF, even without overt ACS, may improve clinical outcomes.

In conclusion, Velazquez sees evidence in favor of revascularization for HFrEF in observational studies and RCTs. Moreover, no groups that are harmed by the procedures have been identified in RCT subgroup analyses, nor is there an age cut-off, an there is no imaging pre-requisite to perform CABG. Thus, he advocates – as does dr Cleland he repeatedly pointed out – for revascularization in HFrEF, provided we look for the right patients.

In a brief rebuttal, Cleland expressed his surprise that Velazquez could not think of another therapy that provided that mortality benefit; how about beta-blockers, MRA, sacubitril/valsartan and more? He also pointed out that Velazquez showed CV mortality rather than all-cause mortality, while with aging, other causes come into play. As often in these debates, Velazquez admitted that they agree on most things. What is clear is that we should use all available evidence. We should be looking for these patients, and we should treat them as we understand is good. This means that it is not appropriate for every patient, but many may benefit.