Aspirin Advocate vs Skeptic: Any Common Ground?

Paul M. Ridker, MD, MPH; John G.F. Cleland, MD


October 29, 2014

This feature requires the newest version of Flash. You can download it here.

Adversaries on a Controversial Topic

Paul M. Ridker, MD, MPH: Welcome to on Medscape Cardiology. I am Paul Ridker, a cardiologist at the Brigham and Women's Hospital in Boston and the Harvard Medical School. With me here is John Cleland from the Imperial College and the Royal Brompton Hospital in London.

We are going to discuss aspirin, an interesting and controversial topic. I was involved in some aspirin research earlier in my career. I would like to divide the aspirin controversy into several parts, starting with unstable angina and acute coronary ischemia, and then secondary prevention, and ultimately primary prevention, where the real controversy exists.

You have written some very controversial articles about this. Controversy is a good thing. It makes us think and challenge dogma and what we know.

John G.F. Cleland, MD: And if I'm wrong, somebody will prove it.

Dr Ridker: That's right. Let's start with acute ischemia/unstable angina. From my perspective, we start with ISIS-2,[1] a very large randomized trial with a factorial design. Did aspirin work or not?

Dr Cleland: It worked. There are very few holes in the ISIS-2 study. If we started to criticize it, we would seem petty. This was a positive study. We should remember exactly what ISIS-2 delivered.

ISIS-2 was 162.5 mg aspirin daily, not 75 mg daily. I don't know why the cardiology community in Europe came up with this odd number. Treatment was given for 28 days, and the primary endpoint was all-cause mortality.

For 28 days of treatment, there was a benefit of 35 days. There was no harm from withdrawal of aspirin for those 7 days. They could demonstrate that 10 years later, there was still a benefit from taking aspirin. The disease was modified in a similar way that thrombolysis modifies the disease by taking aspirin for 28 days. That's very important. Why take aspirin every day for the rest of your life, but we don't give streptokinase for the rest of your life?

Dr Ridker: Let's be careful there. We may end up agreeing, at least on this.

Dr Cleland: On dose and duration?

Dr Ridker: Right. ISIS-2 was a 2 x 2 factorial study with streptokinase at a time when thrombolytic therapy was very controversial; it hadn't been tested yet. We have to give the ISIS investigators phenomenal credit for having the guts to test thrombolytic therapy.

What we learned from that trial was that streptokinase alone is very effective, aspirin alone is very effective, and the combination is even better. That was a landmark trial, and it sounds as though we agree on that.

Dr Cleland: It had more than 20,000 patients.

Dr Ridker: Our basis of beginning is 162 mg of aspirin for at least 28 days. In acute ischemia, we would both say, "Keep giving aspirin."

Dr Cleland: I would like to make two more points. One is about the ISIS-1 study,[2] which was completed before ISIS-2. ISIS-1 tested intravenous atenolol vs placebo. The proportion of patients who were discharged on an antiplatelet agent in ISIS-1 was 5%. In ISIS-2, it was a double-blind, placebo-controlled trial in which both treatments were withdrawn for 7 days before the primary endpoint. How many of those patients would have been restarted on aspirin on the basis of no evidence whatsoever?

Dr Ridker: Saying "Would you give streptokinase every day the rest of your life?" is misleading, because that is very aggressive thrombolytic therapy. It's not like aspirin. The risk-to-benefit ratio is part of this discussion. In acute ischemia, we are on the same page.

Secondary Prevention: Smoke and Mirrors?

Dr Ridker: Let's go to the next level, where we might or might not be in agreement. This is chronic, long-term secondary prevention. We are now into meta-analyses, because we have many trials.[3] There are different ways of looking at the data.

Dr Cleland: What drives the meta-analysis? The positive result of meta-analysis is driven by small clinical trials, because the large ones were all neutral.

There is incredible publication bias in those secondary prevention studies, because there are no negative small studies. There are only small positive studies, but no large positive studies. They are all neutral. The whole thing is smoke and mirrors driven by small studies, and some of them had impossible results in terms of the efficacy of aspirin. They either weren't very well controlled, or they were very lucky studies.

Dr Ridker: I'm hearing smoke, mirrors, and luck in the setting of randomized trials. We do randomized trials to avoid the smoke, mirrors, and luck. I don't understand why these words are being used to describe these studies.

Dr Cleland: The meta-analysis is the smoke and mirrors. The large studies, such as the PARIS studies[4,5] in acute myocardial infarction and the AMIS study[6]—these three large studies, PARIS I, PARIS II, and AMIS, were all neutral. Those studies used 1 g of aspirin daily. Some people said that was too much aspirin, but you can't argue a positive by saying, "We gave the wrong dose, so we imagine that the right dose might have worked, but we don't actually have any evidence."

Dr Ridker: I will grant you that one of the complexities of this field that has allowed this debate to go on so long, and why we are having new trials, is because dosing has been inconsistent from study to study and setting to setting.

Given the overwhelming benefit in acute ischemia—we agree on that—and given a meta-analysis that sounds quite promising (roughly 20% reduction in the major cardiovascular events), I'm struggling with the hesitancy to acknowledge that these patients are benefiting. They are at very high risk. Why wouldn't we treat them?

Dr Cleland: Your 20% benefit is predicated on the small, impossibly positive studies that have been published, and the lack of publication of small, impossibly negative studies. There is a publication bias.

Dr Ridker: Are you asserting that there are small unpublished studies, or do you know that there are small unpublished studies? There is a big difference.

Dr Cleland: I don't know.

Dr Ridker: I just want to establish that you are not saying there are unpublished data; you are just hypothesizing there might be.

Dr Cleland: Yes.

Dr Ridker: As an investigator, most of us do trials because we are genuinely interested in trying to understand what is good what is bad for our patients. Most of us begin trials accepting the fact that what we are testing might work or might not work. We have hypotheses that have failed, and some that have succeeded.

I don't see a "big bad evil" element in the aspirin story. There isn't much money involved here (it's generic). You are not saying there are missing data. You are just suggesting that there might be.

Dr Cleland: If you do publication bias plots (funnel plots), you expect symmetry around the size of the study and the eventual meta-analytic benefit of aspirin, but there isn't. The large trials are to the neutral side of the meta-analytic expectation that is being driven by publication bias.

Dr Ridker: I am going to continue giving aspirin in secondary prevention. My sense of the same database is that the 20% relative risk reduction is probably worth it.

The original intent of a meta-analysis was to estimate sample size for a new study. But in this case, we are doing it because of the absolute size of the secondary prevention trials—many were small in an era where we weren't doing mega-trials yet.

Again, I want to give the Oxford group credit for taking us to those mega-trials, but my clinical sense and my reading of this is that the risk of these patients is quite high. It would be very difficult in 2014 to placebo-control that setting.

Primary Prevention: Amp Up the Debate

Dr Ridker: Let's move on to primary prevention, where you have been very vocal. It is much more controversial, because even aspirin advocates might have some frisson here, so to speak.

Charlie Hennekens was the scientific mentor who taught me epidemiology. I worked with him on the Physicians' Health Study. I am the first author of the Women's Health Study paper,[7] but I have to admit that in that setting, the benefits were small. They were real, but they were small, and limited to a subgroup who looked to be very high-risk.

I am also principal investigator of a big statin trial (the JUPITER trial[8]), and I'm very cognizant that the aspirin data largely precede the statin data. Is that part of the difficulty for you, that we weren't giving statins at the time?

Dr Cleland: Yes, but not a major part of the difficulty. We have robust evidence from statins, and we have robust evidence that treating hypertension reduces risk, so we can modify the risk for stroke and myocardial infarction effectively through two mechanisms. The question is what, if anything, aspirin adds to that, or whether it detracts.

One reason we are doing these large trials is to reduce disability and death. Another controversy I will throw at you is that stroke and myocardial infarction are surrogate outcomes for disability and death. The disability and death that they cause are what interest me, not the events. We don't have any evidence that aspirin reduces disability and death for primary prevention. Disability and death are the ultimate expression of these diseases. In the United States Physicians' Health Study,[9] there was one cardiovascular death difference between aspirin and placebo groups after 100,000 person-years of follow-up.

Dr Ridker: What I hear you saying is when the US Food and Drug Administration suggested that the data on primary prevention were somewhat equivocal; you would not disagree with their assessment.

Dr Cleland: They were a bit slow in the uptake, but I welcome the fact that they finally got there.

Dr Ridker: There are ongoing aspirin trials in primary prevention—two in diabetes,[10,11] two in moderate-to-high-risk patients,[12,13] ? They will involve about 45,000-55,000 patients. Hopefully, that will be enough to settle the discussion on the hard outcomes.

Let's Agree to Disagree

Dr Cleland: Absolutely. I look forward to that.

There is another aspect to this that we should think about. Why do cardiologists use aspirin? It's the idea of vascular disease as a thrombotic disease—that the platelets are sticky, and when there is plaque rupture you need to stop the thrombus propagating, and aspirin can reduce that. That is the philosophy. When you look at these plaques, there is a lot of hemorrhage. There is a lot of hemosiderin.

Dr Ridker: This is fascinating, because it is one of those areas where normally we have a biological construct. Then we do trials, and we have to change our construct. I am hearing that this is the construct from a while ago. I don't agree or disagree. We are trying to figure out how to do these trials.

Dr Cleland: In some people, there is rupture of the fibrous cap and they have a thrombotic incident. In other people, the capillary coming from the vasa vasora ruptures and bleeds into the plaque. Bleeding is the primary event. It may be that both are operating.

In stable patients, one half of the vascular events are bleeding events and the other half are thrombotic events. You give a drug that does both things, and it does nothing. Have I convinced you?

Dr Ridker: No. What I would argue is that we are going to agree on some things and agree to disagree on others. We agree on acute ischemia. On secondary prevention, I am going to treat. You are probably not going to treat, but I'm going to treat because they are very high-risk. I don't know how the trial can be redone. In primary prevention, I am going to give my patients a statin first.

Dr Cleland: I would treat the hypertension.

Dr Ridker: They are being treated for their hypertension. They are counseled about diet and exercise. They are not smoking. My first therapeutic prescription is going to be for a statin, and I currently limit my aspirin use to people who are already on a statin and are particularly high-risk because I think that they have atherosclerosis, trying to get that number needed to treat down. We are both going to look forward to the trials, which should be completed in a few years.

This is Paul Ridker. My guest has been John Cleland. Thank you for joining us for this controversial topic on Medscape for


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.