GRAND DEBATE: Inflammation and cardiovascular disease
ESC 2017 - Barcelona
A Grand Debate of course focused on inflammation, to compare differing interpretations of the CANTOS trial data. Paul Ridker and Alberico Catapano debated how central inflammation is to CVD. This is an impression of this Grand Debate held at ESC 2017.
Positioning of the topic - Inflammation and cardiovascular disease
Thomas Felix LUSCHER (Zurich, Switzerland)
Thomas Lüscher went back in time to illustrate that the concept of a role for inflammation in atherosclerosis has been around for a long time, which dates back to 1859 when R. Virchow wrote that “an inflammation of the inner arterial coat is the starting point of the so-called atheromatous degeneration.”
We now know a bit more about the mechanism of plaque build-up, and we can look at it with imaging techniques. Imaging has revealed that not all plaques are composed of the same material; the composition affects the characteristics of the plaque and possibly whether a plaque is stable or prone to rupture. The question rises whether a link exists between inflammation and plaque rupture.
Various clinical symptoms such as multiple sclerosis, rheumatoid arthritis, ulcerative colitis and scleroderma, are linked to inflammation. Some of these symptoms have been linked to endothelial function. This can be assessed through measuring dilatation in response to nitric oxide and shear stress with flow-mediated dilation. Indeed, it has been shown that inhibition of IL-1 with anakinra improved vascular (and left ventricular) function in patients with rheumatoid arthritis.
Some cytokine levels are increased locally at the site of plaque rupture and plaques are full of inflamed cells. Activation of TLR4 and to a lesser extent TLR2 at the site of the thrombus suggest that native immunity is involved locally. Inflammatory cytokines in turn induce tissue factor expression. TNF-α blockade in patients with rheumatoid arthritis was shown to improve endothelial function over the course of several weeks, but not all inflammation-lowering strategies were effective. So, in addition to the known residual cholesterol risk, the CANTOS results reinforce the proposed concept of residual inflammatory risk.
Inflammation plays a key role in coronary artery disease – PRO
Paul M RIDKER (Boston, MA, USA)
Not surprisingly, Paul Ridker, as principal investigator of CANTOS, got to defend the PRO-position. While the concept of inflammation has been appreciated for a while, the tension until yesterday was there was not therapy to offer, said Ridker. Nice articles have been written about the cellular processes in inflammation, but he wanted to focus on the clinical aspects.
In numerous studies, inflammation has been shown to be a strong and consistent predictor of CV risk. The exploration of inflammation as a therapeutic target started with a statin trial, since statins have some anti-inflammatory properties. Actually, the lipid-lowering and anti-inflammatory effects of statins are very intertwined, but efforts have been made to look at the phenomena separately.
In PROVE-IT and IMPROVE-IT for instance, it became clear that patients who achieved both an LDL-c and a hsCRP goal were best of in terms of recurrent vascular events, as compared with either goal alone or attaining none. Thus, getting both LDL and CRP down is the pathway towards the greatest benefit for patients.
The Jupiter trial results in patients with LDL-c below guideline targets but with elevated CRP, changed the guidelines and clinical practice. It did, however, not answer the question of the current debate.
The Emerging Risk Factor Collaborators described that the magnitude of independent risk associated with inflammation is at least as large as, if not larger than that of blood pressure and cholesterol.
When discussing the CANTOS results, it is important to realize that there are many other potential pathways for potential benefit. The CANTOS researchers first chose the IL-1 - IL-6 pathway based on available data, but this is only the beginning of a series of trials that explore various relevant targets. After the CANTOS study had been initiated, Mendelian randomization results corroborated with a strong signal that the researchers were looking in the right direction.
Naturally, dealing with inflammation means dealing with an intricate balance; inflammatory processes remain important in many situations. This led to the idea to only interfere with IL-1β, leaving IL-1α unaffected, hoping to thereby limit side effects. IL-1β is part of the NLRP3 inflammasome. The NRLP3 inflammasome interacts with cholesterol crystals to induce inflammation. Thus, IL-1β seemed a good target to inhibit, hoping to thereby reduce inflammation, and ultimately yield clinical benefit. Canakinumab is a monoclonal antibody directed at IL-1β, which is injected every 3-4 months. Three doses were tested in the outcomes study, in patients with persistently elevated hsCRP: >2 mg/L (average: 4 mg/L). Of note: in studies with patients with residual cholesterol risk, the average hsCRP levels is 1 mg/L. Average LDL-c level was lower in CANTOS than for instance in the PCSK9 inhibition FOURIER trial. No change in lipid parameters was seen in CANTOS, but a large reduction in hsCRP and IL-6.
The detailed CANTOS results can be read elsewhere on this site, but all endpoints were reduced in the canakinumab-treated group, although not all statistically significantly. The primary MACE endpoint was significantly reduced by 15% with the two highest tested doses of IL-1β-inhibition. No effect modification was observed. Interestingly, the effect was larger (27%) in those with reductions in hsCRP >median at 3 months (1.8 mg/L), while no effect was seen in those who did not show hsCRP reduction. Ridker shared that yesterday he was asked why no number needed to treat was calculated. His answer is that he would not treat those that do not respond. He proposes to start treatment and measure hsCRP after 3 months, and only continue treatment in those who show a response to therapy.
A very clear dose-dependent decrease in all-cause cancer mortality was also observed, which was almost entirely driven by a dose-dependent reduction in lung cancer. In the first 2 years of canakinumab treatment, no fatal lung fatality was observed. These findings need to be replicated, but they are very exciting.
Ridker concluded by saying that we now have at least one agent proven to be effective. In PROVE-IT and IMPROVE-IT about 30% had residual inflammatory risk, and about 14% had both residual cholesterol and inflammatory risk. This suggests that if you are not measuring CRP, you are missing something. Lower is better appears to be true for both LDL-c and hsCRP, both in primary and secondary prevention.
Inflammation plays a key role in coronary artery disease – CON
Alberico CATAPANO (Milan, Italy)
Catapano started out by admitting that he wondered why he would defend the contra position in this debate, since he well knows that inflammation is there and that it is important. He does have a few nuances to share, though, and aimed to pinpoint what we don’t know. He focused on individual susceptibility, since in studies we tend to look at averages. He suspects that some patients are hyper-responders. Also, he emphasized that inflammation is not there until there is a trigger. Cholesterol crystals can be an important trigger of inflammation and the NLRP3 inflammasome is a common node in complex inflammatory diseases. But how is inflammation induced specifically in the arterial wall?
Catapano looked at the Hill criteria for causality for LDL: all boxes are checked. Does the same apply to inflammation? He concluded that we are not at the same point yet; we need more information and confirmation. A very consistent relation exists between LDL and CVD and evidence stems from different studies. Moreover, the longer you are exposed to lower levels, the higher the benefit. LDL is a good marker because it is involved at every stage of atherosclerotic plaque formation. Inflammation on the other hand comes and goes.
Catapano therefore argued that LDL is what leads to inflammation. Some patients are more susceptible to triggers of inflammation. It is an open question to address in the future what these triggers are exactly. Catapano expected Ridker to say that he wants to investigate what is behind the response to canakinumab therapy. Catapano postulated that LDL may be very good candidate for that. He points to the similar event reductions in CANTOS as compared with FOURIER, the PCSK9 inhibition trial, to underscore his point.
Catapano thinks of LDL as a non-resolving inflammation; the stimulus is not really removed with as a consequence a low degree of inflammation. He expects that Ridker is looking at patients who are most sensitive to low LDL-c levels. This needs more study. Is it possible to modify this pathway? In experimental models, inhibiting the inflammasome has been shown to inhibit atherosclerotic lesion development. Accumulation of cholesterol in dendritic cells triggers the immune system to develop inflammation: this is a very important pathway that needs clarification.
Catapano concluded that yes, inflammation is there, and it’s important. But if you don’t have the trigger, there is no inflammation. If there is no cholesterol, there is no inflammation.
Discussion with the audience
Ample time was given for the audience to ask questions. One question was whether it would not be better to treat in the beginning of the inflammation process, when oxidation of LDL takes place, rather than in the middle. Indeed when waiting until atherosclerosis is clinically evident, it is already very late. But it is difficult to be so early and an intervention needs to be very very safe. Ridker would say that statins are safe, but others might disagree. Therefore for now we are limited to giving lifestyle advice in an attempt to stop early oxidization.
The downside of reducing inflammation was also discussed. Canakinumab is well-known to give mild reduction in white blood cells. Any infection was not statistically significantly increased, but fatal infections were. Ridker argued that all cardiologists in the room have to become a rheumatologist pretty soon, if they will start using a drug that is used to treat a non-fatal disease. With this he meant to say that rheumatologists have learnt to work with it: they instruct their patients to come to the clinic earlier if fever develops. Physicians bring in patients quicker. Ridker therefore expects that the extra infections will be controllable.
Someone asked whether we should add residual thrombotic risk to the list, considering the COMPASS results. Ridker replied that he was going to mention this in his rebuttal, as a call to keep thinking broadly. He thinks of it as residual remnant risk, and several studies are evaluating the effect of fish oil, or lowering Lp(a) or other lipoprotein particles. Understanding more of their effect will imply more personalized medicine. Ridker said that we have been lucky: statins have been very effective in large groups. Many other interventions will likely be less generally effective. He thinks we have to learn provide more biomarker-driven care. But that requires a lot more research. He emphasized that these data on IL-1β are not generalizable to other targets in the pathway. The effect of every target has to be shown separately. Even when targeting IL-6: although closely related in the pathway, the side-effects may be different from inhibiting IL-1β.
Somebody pointed out that atherosclerosis develops everywhere, but some areas are more prone to plaque formation than others. This can be explained by turbulence. Catapano added that this fits well with inflammation, since you need something to get the LDL there, which then triggers inflammation. Thus perhaps an athero-prone flow that triggers the inflammasome is a mechanism to explain the localized effect.
And will it change practice now, was a question to Ridker, or mainly the beginning of a series of studies? Ridker replied that to his astonishment, ‘we are still debating the safety of statins. That is a failure. The fact that 25 years after the 4S results, we can’t agree as a group, that is a problem.’ As a clinician, he wants to see his patients get better. If that does not work, he will stop using the drug, with the argument that he does not want to risk inducing side-effects and unnecessary costs. Ridker thinks that will be the approach in clinic.
Filippo CREA (Roma, Italy)
In his concluding remarks, Crea first focused on the question whether CRP assessment is clinically important in primary prevention. He showed data of the INTERHEART study that revealed that about 45% of the events is attributable at a population to traditional risk factors. Another 45% is not in traditional risk charts. Can CRP be useful in this regard? CRP correlates with depression, suicidal ideation, obesity, diet and physical activity. So indeed CRP may capture risk that is not contained in traditional risk factors.
Furthermore, he asked whether CRP assessment is clinically important after ACS. Adaptive immunity is profoundly dysregulated in ACS. Indeed, based on data he concluded that in patients with recent ACS, assessment of CRP may identify patients with an inflammatory cause of instability, and this may guide treatment decisions. It should, however, be noted that a large variation is seen in CRP levels: many patients have almost undetectable levels. In the very acute phase of STEMI it has been shown that 41% have <2mg/dL.
So we have to think of alternative hypotheses: we do not know much about cholesterol crystallization in plaques. Different pathogenic processes take place in different types of plaques.
Plaque erosion does not follow the same mechanism as plaque fissure. This means that the processes need to be targeted differently. The process of fissure can be counteracted, but now we need to see how other factors can be affected.