COURAGE: Does 15-Year Data Have Any Clinical Relevance?

pills_vs_stent_140sqThis past week saw publication of an update to the 2007 COURAGE trial which compared optimal medical therapy (OMT) to stenting (PCI) as the initial management strategy for stable coronary artery disease. (Please note the phrase “initial management strategy.” This will not be the last time you see it in this post.)

Appearing in the New England Journal of Medicine and titled “Effect of PCI on Long-Term Survival in Patients with Stable Ischemic Heart Disease,” this study, performed by Dr. Steven P. Sedlis and other COURAGE trial investigators, is described as “an extended survival analysis to examine the potential long-term survival benefit from initial PCI among the patients with stable ischemic heart disease who were followed for up to 15 years after initial enrollment in the COURAGE trial.”

This new analysis concluded: “…we did not find a difference in survival between an initial strategy of PCI plus medical therapy and medical therapy alone in patients with stable ischemic heart disease.” This also was the finding of the original COURAGE trial at five years.

Headlines and tweets from the medical media followed:

But I beg to differ. I think that this 15-year analysis has so many statistical weaknesses, and is looking at a snapshot in time that has long been surpassed by advances in technology and practice, that the conclusion of this “extended survival analysis” is basically irrelevant to current practice.

First, some history. (Of course, you’re welcome to skip the history, but only at the risk of repeating it. H/T G. Santayana.)

History of the COURAGE Study
There was considerable controversy when the COURAGE trial was first published in 2007. Interventional cardiologists pointed out that 90% of the CAD patients had been excluded, that the population studied represented lower-risk cases where a mortality benefit would never have been claimed, that almost all the stents used were the older, less effective, bare metal variety, that one-third of the patients in the OMT-only cohort had crossed over to PCI during the original follow-up period. And so on.

One widespread opinion among interventional cardiologists was simply that COURAGE showed nothing new because guidelines already recommended OMT as the first-line treatment strategy for stable CAD, and that PCI was a way to relieve angina, not to increase survival. This was news, however, to many patients and physicians – several studies showed that patients were under the impression that getting a stent prolonged life. At Angioplasty.Org we tried to clarify some of these issues (see “Answers to Top Ten Questions About Stents and Angioplasty vs. Drug Therapy“).

The COURAGE trial certainly had an impact on clinical practice and, coupled with new concerns over stent thrombosis seen in drug-eluting stents, the use of PCI for stable angina decreased over the next few years.

However, during the past decade, a number of developments have occurred: more specific Appropriate Use Criteria (AUC) have been fashioned, at least two new generations of drug-eluting stents have come to market, with improved outcomes and lower complication rates, and the use of intravascular guidance and imaging, such as fractional flow reserve (FFR) and IVUS or OCT, have grown in utilization.

For example, the FAME and FAME 2 studies clearly showed that PCI, when guided by FFR, targeted blockages that could benefit from revascularization more accurately than angiography alone. The FAME trial resulted in one-third less stents being used and one-third improvement in outcomes (see “Better Outcomes for Stents When Fractional Flow Reserve (FFR) is Used“). In fact FFR is a central aspect of the ISCHEMIA study, which the medical community hopes will address the treatment of stable CAD more accurately than COURAGE. (I hope so, too, although ISCHEMIA is having enrollment problems, as pointed out by Larry Husten in his recent Cardiobrief post.)

Limitations of the 15-Year Analysis
I always find the last few paragraphs of peer-reviewed journal articles most enlightening. That’s where the limitations of a study are usually listed and the 15-year analysis of COURAGE has many. In fact I would suggest that anything more than the original conclusions of the COURAGE trial are not supported and that to extend those conclusions to 15 years distorts the message. Here’s why.

Loss of Follow-Up: 47%
COURAGE enrolled 2,287 patients, some from the Veterans’ Administration system, some from Canada, some from other non-VA sources. The 15-year mortality data could only be tracked via Social Security numbers and, due to privacy issues, the VA patients were virtually the only patients whose Social Security numbers were available. This meant that there was survival data for only 1,211 patients: a 47% loss of follow-up! Moreover, since VA patients tended to be older, suffering from more co-morbidities, and predominately male, the follow-up data were further skewed.

Cross-Over to PCI: Unknown
During the ∼5 year study period of COURAGE, 32.6% of patients in the OMT group crossed over to PCI, presumably to relieve angina or to deal with a more urgent issue. But the outcomes for these patients were still counted in the OMT cohort because, as defined, COURAGE compared OMT to PCI as an initial management strategy. (I told you that term would be mentioned again!) So how many patients crossed-over to PCI (or CABG) during the following decade? We don’t know because there is no data on how many of the OMT cohort progressed to revascularization. All the investigators know is whether these 1,211 patients were still alive or not. Given that a third of OMT patients crossed over to revascularization in the first five years, does anybody want to offer a guess at how many crossed over in the following ten?

What’s the significance? Well, the authors write:

“A high rate of revascularization during the extended follow-up period would also reduce the likelihood of a divergence in mortality between the two study groups; therefore, the late follow-up data may be less reflective of the initial treatment assignment than of a convergence of management strategies over time.”

Or you could say that a high rate of cross-over would basically invalidate the study’s conclusion that there’s no mortality difference between OMT and PCI, but we don’t have any idea what the cross-over rate was, so we’ll just make believe there’s no difference.

Cause of Death: Unknown
One would think that, when trying to compare the mortality associated with two management strategies for coronary artery disease, it might be helpful to know if death was due to cardiac problems, or some other cause, like cancer or a car accident. Again, there is no data regarding cause of death.

Relevance to Today’s Treatment Strategies
The authors also discuss the significant changes in stent design, use of FFR, IVUS and OCT that I discussed above and they stress the importance of the ISCHEMIA trial to resolve these questions for current practice.

*               *               *               *               *

So, in conclusion, I think that the lack of data and the limitations of this long-term analysis of COURAGE patients yield virtually no new findings and, if anything, may obscure important treatment realities. One could hypothesize, for example, that over the period of 15 years, most patients treated with OMT as the initial management strategy will ultimately need to turn to revascularization. But we can’t really know.

One thing we do know is that after 15 years PCI is as safe as medical therapy, and that having a stent placed doesn’t increase your chances of dying. Of course (disclaimer here) I am inclined to say this; after all, our web site is called Angioplasty.Org!!

I welcome your comments.


Filed under Appropriate Use Criteria (AUC), Bare Metal Stents, Clinical Trials / Studies, COURAGE, Drug-Eluting Stents, FAME I / FAME II, FFR, Intravascular Guidance, IVUS, OCT, Optimal Medical Therapy, Stent

3 Responses to COURAGE: Does 15-Year Data Have Any Clinical Relevance?

  1. John

    Interesting article. I wish that the COURAGE Trial addressed the survival rate of patients that received a stent when it wasn’t even necessary, as in my case.

    I recently had a review of my angioplasty/PCI procedure by another cardiologist and I was told that my blockage was actually in the 30-40% range and that I should never have received a stent in the first place, especially since I had no chest pains, even when exercising.

    I’ve yet to meet with the cardiologist that placed the stent to see if he wants to apologize or deny the findings of the other cardiologist. The problem he will have is “Pictures don’t lie!!”

    Great comments on the 15 yr. COURAGE study.


  2. “after 15 years PCI is as safe as medical therapy, and that having a stent placed doesn’t increase your chances of dying” – but this also says that medical therapy is as safe as the new-fangled PCI, which means that people should opt for medical therapy!

    Since all operative procedures have some risk, as does medical therapy, these risk factors should made the difference. But of course there are no recommendations on this: surgeons like to go in there and think they’ve fixed things – they have no patience for slower medical therapy and cautious observation.

    The patients, though, are being pushed into surgery – by semantics, not data.

    • Alicia – Thanks for your comment. As per the professional society guidelines, optimal medical therapy (OMT) is the first line treatment for stable coronary artery disease. But if symptoms remain (as they did in one-third of the COURAGE patients) then PCI is a choice that can be considered. Also, PCI is not surgery. It’s performed by interventional cardiologists. But certainly we (and the societies) agree that a cardiologist should not be performing PCI in a stable angina patient, when the patient has not yet tried OMT.

Leave a Reply

Your email address will not be published. Required fields are marked *