Understanding Cardiovascular Outcomes Trials

Peter L. Salgo, MD: Give us a historic perspective, if you will, on the cardiovascular outcomes trials.

Christian T. Ruff, MD, MPH: It’s interesting. All of the trials and diabetes medications, predominantly, were targeting glycemic surrogates.

Peter L. Salgo, MD: It’s a sugar disease, so let’s work on the sugar.

Christian T. Ruff, MD, MPH: Right. Is it effective in reducing a hemoglobin A1C? And do we detect any adverse events? And then, there was, initially, this whole milieu that kind of arose around the TZDs (thiazolidinediones), and particularly, rosiglitazone, in the early 2000s. There was a seminal meta-analysis that said one of these drugs had increased cardiovascular risk. It actually got a huge amount of publicity. There was Congressional inquiry around it, and the FDA released guidance in 2008 mandating that cardiovascular outcomes be done for all new applications for diabetes medications to specifically address whether these drugs cause cardiovascular harm.

Peter L. Salgo, MD: If I can paraphrase that: the FDA was looking for cardiovascular safety.

Karol E. Watson, MD, PhD, FACC: They said, “We’ll approve it on glycemia.”

Christian T. Ruff, MD, MPH: On glycemia.

Karol E. Watson, MD, PhD, FACC: But, you have to then prove that it’s safe.

Christian T. Ruff, MD, MPH: Yes, in preapproval, there had to be some evidence; it couldn’t be egregiously harmful. Then, they narrowed it down post approval. But you have to show us preapproval and then post approval that these drugs are safe in patients with established cardiovascular disease. So, they’re saying, “You have to do a trial in high-risk patients with cardiovascular disease and show that there’s not an adverse signal, particularly around heart attack and stroke.”

Stephen A. Brunton, MD, FAAFP: This is part of the strategy of the ADA [American Diabetes Association]. They were hoping that this would create a change because we’ve been using these agents without proving not only that they’re not safe, but that they show a benefit. So, the cardiovascular outcomes trials were designed to show noninferiority. And then, if it gets to certain thresholds, superiority. And that was the strategy.

Christian T. Ruff, MD, MPH: I want to make a comment about that because I think it’s important. We now are going to talk about drugs that were studied for cardiovascular safety but, in fact, may have cardiovascular benefit. But when you design a study to address cardiovascular safety (so, a noninferiority study), you are not adequately designing a study to ask, “Is this drug beneficial for heart disease?” Because there’s a lot of therapies for conditions (including blood pressure reduction and lipid management), to do a superiority trial, you have to do a much longer duration trial. So, the studies are not adequately designed to definitively say whether these drugs are beneficial long term.

Peter L. Salgo, MD: That was my question going forward. These are patients, as I understand it, in these studies, that were already patients at high risk. They already had established cardiovascular disease.

Rosemarie Lajara, MD, FACE: Absolutely. Enrich the population with a high-risk population.

Peter L. Salgo, MD: This is not necessarily the best population to prove that there’s an improved outcome.

Stephen A. Brunton, MD, FAAFP: The CANVAS study also had a cohort that was a primary prevention group.

Christian T. Ruff, MD, MPH: Right, about 35%.

Rosemarie Lajara, MD, FACE: One-third.

Peter L. Salgo, MD: But my intuitive guess, though, which I think was wrong, was whether or not you set it up to show a cardiovascular benefit. That might have fallen out of the data, but you’re telling me it didn’t?

Christian T. Ruff, MD, MPH: I’m just saying if you want to definitively ask if a drug is beneficial to reduce cardiovascular events, you would design it differently than you would in a noninferiority study, which is how to say a drug is safe. And so, I will say it’s amazing that we are going to talk about drugs that are beneficial for cardiovascular disease. But if you wanted to study a new, exciting, cardiovascular drug that would lower the risk for heart attack or stroke, you would actually not design it the way these outcomes studies were designed.

Rosemarie Lajara, MD, FACE: Oh, no. The intent is different.

Stephen A. Brunton, MD, FAAFP: The very cost of it, and the size and the length of these studies, really precludes them being done. The offshoot of this is now we have data we never had before. Most people assumed that if you get good blood through a control, you’re going to have cardiovascular benefit. But it hadn’t been shown.

Christian T. Ruff, MD, MPH: Here’s the key thing: if you look at, say, cholesterol-lowering trials or even at the PCSK9 (proprotein convertase subtilisin/kexin type 9) inhibitors, they don’t reduce cardiovascular mortality. And that’s because it’s very hard, when you’re dealing with atherosclerotic events, to reduce fatal events and stroke in short duration trials that have an optimal therapy. And so, even if these diabetes drugs could have a beneficial atherosclerotic effect, you wouldn’t see it in the duration, necessarily, of the trials for these cardiovascular outcomes trials. I think it’s great that we do have some exciting data, but I think it’s possible that if these trials were done differently or are done differently in the future, there may be even better options.

Peter L. Salgo, MD: But let me just establish the baseline. These initial trials were done because there was some concern that the drugs might increase cardiovascular risk and cause harm.

Christian T. Ruff, MD, MPH: Harm. Right.

Karol E. Watson, MD, PhD, FACC: Harm.

Peter L. Salgo, MD: That’s what they were looking for and they found no change.

Christian T. Ruff, MD, MPH: Actually, none of the trials that have been reported after this FDA guidance actually did show cardiovascular harm.

Peter L. Salgo, MD: But they didn’t show a benefit. And if I understand you correctly, that’s because they weren’t set up to look for it.

Rosemarie Lajara, MD, FACE: Right. That was not the intent of the design, nor the aim of the study.

Transcript edited for clarity.