And you measure genetic polymorphisms?

Correct. In this study, we created a score of 46 different polymorphisms associated with lower LDL, and a separate SBP genetic score consisting of 33 polymorphisms associated with lower SBP. These scores were used as instruments to naturally randomize people to higher or lower LDL, higher or lower SBP, or both, in a factorial design, in exactly the way we would in a randomized trial.

How did you lower blood pressure (BP) and lower LDL?

We didn’t intervene to lower BP. It was the scores that randomized people to higher or lower levels. This is referred to as the concept of Mendelian randomization. Perhaps the easiest way to explain that concept is by way of analogy with a randomized trial.

There are numerous polymorphisms, for example, that are associated with lower LDL, and each of those polymorphisms are inherited approximately randomly at the time of conception in a process sometimes referred to as Mendelian randomization. Inheriting an allele associated with lower LDL is analogous to being randomly allocated to an LDL-lowering therapy, while inheriting the other allele is analogous to being randomly allocated to usual care. If allocation is really random, the only difference will be the LDL level. Then we can follow people forward to get an unconfounded estimate of the causal effect of LDL in a manner analogous to a long-term randomized trial.

All randomized trials I know of have entry and exclusion criteria. Were there any entry or exclusion criteria in this methodology?

There’s a fundamental difference between a person enrolled in a naturally (Mendelian) randomized trial and one enrolled in a randomized trial.

Unfortunately, people enrolled in a randomized trial, for convenience, are those with advanced atherosclerotic disease, because the expectation is they will have the greatest number of events in the shortest period to minimize the sample size. Attempts to lower LDL or BP in this population may have minimal impact (or at least less than optimal impact), because of the advanced atherosclerotic and arteriosclerotic burden. By contrast, the approach of the Mendelian randomization contemplates keeping BP and LDL lower beginning much earlier in the atherosclerotic disease process. It therefore examines explicitly what could happen if we began to intervene on risk factors – not necessarily pharmacologically, but with lifestyle factors and other methods – much earlier in the atherosclerotic disease process.

To prevent a cardiovascular event, it’s got to start in childhood, by preventing risk factors from evolving.

Right. I think this is what we tried to asses in this study. I’m not so sure it necessarily has to be childhood. I think in childhood it’s the perfect time to begin introducing healthy lifestyles. However, whether or not we aggressively seek to set targets or lower risk factors that young is hard to say. I think for LDL certainly it’s a little bit different, because it’s fairly stable over the life course. But BP is slightly different; it remains stable into the 20s and 30s, and when it hits an inflection point it begins to rise linearly with age. The key is whether we can identify when BP begins to rise and prevent its rise with age. This study addresses that issue.

What did you find?

The purpose of this study was specifically to evaluate the causal effect of combined exposure to both lower LDL and lower SBP on the risk of cardiovascular disease. We sought to address this topic because previous observational studies have suggested that persons who can maintain ideal risk factor profiles throughout the whole of their lifetime have a very low lifetime risk of cardiovascular disease. Unfortunately, fewer than five percent of the population can maintain ideal risk factors.

By contrast, previous Mendelian randomization studies from our group and others have suggested that both LDL and SBP are not only causally related to cardiovascular disease, but they have a cumulative effect over time. This would suggest a simple strategy that promotes long-term exposure to the combination of both lower LDL and lower SBP may be effective for preventing cardiovascular disease.

Unfortunately, the causal effect of exposure to both lower LDL and lower SBP on events is unknown. That uncertainty was recently increased after the landmark HOPE-3 trial. As you know, HOPE-3 reported that the combination of lipid-lowering therapy plus a modest blood pressure-lowering regimen didn’t reduce cardiovascular events significantly more than lipid-lowering alone. That has caused some people to question this synergy between LDL and BP on the risk of cardiovascular disease, particularly among persons without hypertension. We specifically wanted to determine the causal effect of combined exposure to lower LDL and lower BP.

What was the average age of patients? What did you consider hypertension and hyperlipidemia?

We didn’t define hyperlipidemia or hypertension as a criterion. In fact, I would go so far as to suggest that the “at the moment” way we treat hypertension is slightly backward. Rather than letting arteriosclerotic injury advance over multiple years to prove the point when the person develops a BP of 150 or 160, and then argue about how low we should lower it, we should probably begin to think about the notion of preventing the age-related rise. This study gets to that.

The average age was 60.7 years, the average LDL was about 3.4 mmol (128 mg/dl), and the average BP was about 126 mm Hg.

So it was relatively low blood pressure.

Yes, relatively normal levels. The LDL genetic scores and the SBP scores were used to naturally randomize people into four groups: a reference group, a lower LDL group, a lower SBP group, or both. We found that the combination of exposure to both lower LDL and lower SBP was associated with a significantly greater benefit than lower LDL alone or lower SBP alone. In fact, the effects were independent, multiplicative, and cumulative over time […] and that effect was quite consistent across multiple different composite cardiovascular outcomes and the individual components of the composite, including myocardial infarction and coronary heart disease death. There was an 84 percent reduction in coronary heart disease death, which translated into a smaller but still significant and robust 36 percent reduction in all-cause mortality. In addition, the effect seems to be consistent across all the subgroups we studied.

So the take-home message is that lowering SBP and LDL each lower risk, and lowering the combination lowers risk a bit better than individually. There is also a cumulative effect over time that’s persistent. Anything else?

The last point is salient. If we’d like to prevent cardiovascular disease, we might have to start slightly earlier than we do now.

Editor’s Note: This interview has been edited for print from a transcript.

References

Ference BA. A naturally randomized trial comparing the effect of long-term exposure to lower LDL-C, lower SBP, or both on the risk of cardiovascular disease. Presented at the European Society of Cardiology Congress, Rome, Italy. August 29, 2016.

Ference BA, Julius S, Mahajan N, Levy PD, Williams KA Sr, Flack JM. Hypertension 2014;63:1182-8.

Ference BA, Majeed F, Penumetcha R, Flack JM, Brook RD. J Am Coll Cardiol 2015;65:1552-61.

Ference BA, Robinson JG, Brook RD, et al. N Engl J Med 2016;375:2144-53.

<<< Return to top

Dr. Wenger: Let us start, Dr. Gibson, by talking about the limitations of aspirin as an antiplatelet agent.

Dr. Gibson: First, aspirin is a good drug. It’s been around quite a while, but it has some limitations. Obviously, it’s associated with gastric erosion and gastrointestinal bleeding. Prostaglandins protect your stomach lining, and by inhibiting prostaglandin production, it causes gastric irritation and bleeding. That’s well known. The other problem is at high doses, because it blocks prostaglandins (which provide some vasodilation), it can be a bit of a vasoconstrictor.

One of the keys to aspirin is that for the right dose, we’ve come to realize that less is more. Just like every antiplatelet and antithrombin, there’s a U-shaped relationship. The optimal dose looks to be down around 75 mg. Any less than that and you lose antithrombotic properties; any more then the bleeding can start.

Less is more. In trials like PLATO, we’ve seen that dropping the aspirin dose led to even better effectiveness of a drug like ticagrelor. Do the thienopyridines need aspirin? Well, back when people used clopidogrel extensively, there were some non-responders. Thank goodness we had aspirin, because if you were a non-responder, then the aspirin was a Band-Aid for the clopidogrel. There are some people who are non-responders to aspirin, and thank goodness, they’re all on a thienopyridine, because that acts as a Band-Aid for the aspirin. It’s a little bit like that line in the Tom Cruise movie, “You complete me.” These two together were offsetting some of the hazards or limitations of each other. Although with some of the newer thienopyridines (like prasugrel and ticagrelor), they’re probably effective on their own without aspirin. When you do all the sophisticated platelet testing, aspirin doesn’t add much.

What are the hints that the thienopyridines might be effective without the addition of aspirin?

The first hint came with the WOEST study, not in acute coronary syndromes (ACS), but in the atrial fibrillation (AFib)/ACS overlap syndrome. They took the very brave step of getting rid of aspirin. In a modest-sized study (600 patients), the risk of bleeding was reduced, cut in half from about 40 percent down to 20 percent. No surprise there. The big surprise came with the effectiveness results: taking away aspirin was just as good with respect to stent thrombosis and death. In fact, they were a little better when you took aspirin away.

"One of the keys to aspirin is that for the right dose, we’ve come to realize that less is more.” — C. Michael Gibson, MD, FACC

Why is that? We’ve come to an increasing realization that bleeding itself may not be bad; but, it’s what it does to patient compliance with all the evidence-based medicine. That was a big wakeup call.

Let’s talk about just the patients anticoagulated for AFib.

With those patients, I think we’re seeing more and more people dropping aspirin and going with an oral anticoagulant like warfarin plus a drug like clopidogrel. We’re seeing that some of the drugs like prasugrel and ticagrelor have not been as well tested in these kind of combinations, so you should exercise a lot of caution. […] What we need are these randomized trials to guide us. If I have one message for our audience today, it’s this: Be very, very careful with full doses of novel oral anticoagulants (NOACs), with aspirin, and particularly with prasugrel and ticagrelor. You will have bleeding.

I think the ongoing trials will help us to find the right dose for NOACs. It will probably be a reduced dose, along with these other drugs. There is a trial of rivaroxaban called PIONEER AF-PCI, that I’m leading, and it is testing three strategies: 1) a WOEST-like strategy where we drop the aspirin and use rivaroxaban at a 15 mg dose, along with clopidogrel or ticagrelor or prasugrel; 2) an ATLAS-like dose, referring to a trial I’d led where we gave baby doses of rivaroxaban 2.5 mg twice a day, along with aspirin plus a thienopyridine; triple therapy, but with baby doses of the anticoagulant; and 3) traditional triple therapy: warfarin, aspirin, and a thienopyridine. We hope to present this trial later this year.

Another trial looks at two doses of dabigatran combined with a thienopyridine; get rid of aspirin. The RE-DUAL trial has dabigatran with a thienopyridine versus warfarin and triple therapy. Then there’s another new trial called AUGUSTUS, which looks at apixaban, and will then randomize people to apixaban or no apixaban in this same setting.

There’s a global leader study that is comparing traditional DAPT with aspirin and ticagrelor to ticagrelor monotherapy, dropping aspirin pretty early on. That’s very exciting – 17,000 patients, big study. There’s another study called TWILIGHT which gives three months of DAPT (aspirin plus ticagrelor). After three months, though, you drop the aspirin in one arm or continue the DAPT in the other arm.

And finally, there is a trial that I’m leading along with E. Magnus Ohman, MD, called GEMINI. In this trial, the thienopyridine is selected, and then either clopidogrel or ticagrelor, and then there is randomization to aspirin or rivaroxaban. We’ll see which one of those two, when added to a thienopyridine, is better. It’s an exciting time, trying to get the right dose of the right drugs to the right patient.

We easily have three years before we get even the first of those reports coming out. You are the clinician. You are facing the patient across the desk in the office. What do you do for your patient with AFib first?

We’ll have the results of PIONEER AF-PCI at AHA this year. We’ll have some information to guide us. Right now, though, personally, I still use warfarin. I use reduced-dose warfarin to get just above 2. I try and get people off DAPT if they have a stent as soon as possible. I try and put in a newer generation of stent that allows me to get them off those drugs by, say, three months. Minimize the duration of triple therapy, get them onto warfarin – either warfarin with some aspirin for a while or warfarin plus clopidogrel for a while, then on warfarin permanently.

We have all the new guidelines that have just appeared on the duration of DAPT for everyone.

Less is more, shorter is more.

The fascinating part about it – and I think this is what people mean when they talk about lifelong learning – is that what we were doing as standard treatment two years ago is outdated, and a year ago may be outdated by the time we get the next major trials reported. Any final words of wisdom for our audience?

Again, be careful with full doses of those new NOACs with aspirin and the thienopyridines. You’re going to have to be very thoughtful. You’re going to have to cut the dose back a bit if you’re using that combination to make sure you don’t have excess bleeding. Be very careful.

Editor’s Note: This interview was edited for print from a transcript.

<<< Return to top