William E. Boden, MD, on A New Risk Score for Mortality and Myocardial Infarction in Patients With Stable Coronary Artery Disease: The COURAGE Risk Prediction Tool

Using data from the COURAGE Trial, researchers recently developed a risk prediction tool for assessing the probability of mortality or myocardial infarction among patients with stable coronary artery disease. According to William Boden, MD, the COURAGE Risk Prediction Tool is one of the most important developments to come from the COURAGE Trial. In this video, Dr Boden explains how the tool works and what its impact can be on treatment optimization. 

Additional Resources:

  1. Boden WE, Hartigan PM, Mancini GBJ, et al; COURAGE Trial investigators. Risk prediction tool for assessing the probability of death or myocardial infarction in patients with stable coronary artery disease. Am J Cardiol. Published June 7, 2020. doi:10.1016/j.amjcard.2020.05.046
  2. Boden WE, O'Rourke RA, Teo KK, et al; COURAGE Trial research group. Optimal medical therapy with or without PCI for stable coronary disease. N Engl J Med. 2007;356:1503-1516. doi:10.1056/NEJMoa070829
  3. Antman EM, Cohen M, Bernink PJ, et al. The TIMI risk score for unstable angina/non-ST elevation MI: a method for prognostication and therapeutic decision making. JAMA. 2000;284(7):835-842. doi:10.1001/jama.284.7.835
  4. Fox KA, Dabbous OH, Goldberg RJ, et al. Prediction of risk of death and myocardial infarction in the six months after presentation with acute coronary syndrome: prospective multinational observational study (GRACE). BMJ. 2006;333(7578):1091. doi:10.1136/bmj.38985.646481.5 
  5. Bhatt DL, Steg PG, Miller M, et al. Cardiovascular risk reduction with icosapent ethyl for hypertriglyceridemia. N Engl J Med. 2019;380:11-22. doi:10.1056/NEJMoa1812792

For more videos, visit the Resource Center.

William E. Boden, MD, is a professor of medicine at Boston University School of Medicine and a lecturer of medicine at Harvard Medical School. He is also the scientific director of the Clinical Trials Network at VA Boston Healthcare System.


William Boden: I want to thank everyone who’s joining us today to hear about this recent publication in the American Journal of Cardiology that’s predicated on our COURAGE Trial Risk Score. 

The COURAGE trial, I think as most people know, was published in 2007 in the New England Journal of Medicine. It was a large, randomized trial of almost 2300 patients with stable ischemic heart disease who had chronic angina for many years. And they were randomized to a strategy of PCI plus optimal medical therapy—or OMT for short—vs OMT alone, so medical therapy alone, followed for 5 years. The primary endpoint for COURAGE was death or MI over a median 4.6-year follow up.

This was a large, landmark trial that I think has been widely accepted in the cardiology and internal medicine community. And from the COURAGE data set, we thought it would be appropriate to model a new risk prediction tool or what we call the COURAGE Risk Score from the data contained in the trial.

Now, why do we need a risk score in patients with stable ischemic heart disease? Well, the simple answer to that is, we’ve never had a risk score for patients with chronic stable angina or stable coronary artery disease. I think most individuals—both in cardiology internal medicine and so forth—recognize that we have had several very well-validated and used risk scores for patients with acute coronary syndrome and these notably include the TIMI risk score that was developed at Harvard and the Brigham back in the late 1990s. And then shortly thereafter there was a Grace risk score that was developed in collaboration with our colleagues from the UK and the United States. So both the TIMI rescore and the Grace risk score were essentially created because of the fact that we recognize that patients with unstable action or acute coronary syndrome, you know, can be risk stratified into low, intermediate, and high-risk subsets.

Following that basic premise, that patients with stable ischemic heart disease should likewise be able to be risk stratified, we sought therefore to develop and model a risk or from the data derived from the COURAGE trial. And again, we explored 23 candidate variables from the baseline characteristics and laboratory tests and so forth. And of those 23 original baseline characteristics, 9 variables or covariates emerged as being independently predictive of the risk of developing death or MI over a 5-year follow-up period.

The beauty of the COURAGE risk score is that, number one, it’s derived completely from a population of patients with stable ischemic heart disease. Number two, it looks at the outcome of death or MI, which was the courage primary endpoint over a 5-year follow up. And again, it is modeled to hopefully dispel the myth that all patients with stable ischemic heart disease are uniformly low risk, which obviously to my way of thinking has never been intuitive, but many people persist in that belief. 

So of the 23 candidate variables, I already alluded to the fact that 9 of them turned out to be independent predictors. And so the way the risk score is set up is that these are the following risk factors or the variables, the covariates, that went into the independent prediction model. So, heart failure gives you 3 points. In fact, this was the most powerful predictor of outcomes from the COURAGE risk score. And it’s interesting that heart failure would be so prominent because, again, this was a study of chronic angina. And actually, the mean ejection fraction was 60%, so these are patients who almost by definition have preserved left ventricular function. And despite that, heart failure was a powerful predictor of death or MI. So 3 points for heart failure. And then for the number of vessels involved—so 1-, 2-, or 3-vessel disease, you get a point for each vessel involved. So 3-vessel disease would be 3 points. Age gives you 1 point for each 15 years above the age of 45. So a patient, for example, who was 75, that would give you 3 points. Diabetes is a point, prior MI is 1 point, prior revascularization is 1 point, female sex is 1 point, current smoking is 1 point, and then interestingly HDL— low HDL below 30—is 2 points, and HDL between 31 and 40 is 1 point. Now, many people may say, “Wow, why is it LDL in the model?” And you know we look at this very, very carefully, but because these patients were so intensively treated with statins and ezetimibe and other dyslipidemic agents, I think this no longer became a predictive variable. So HDL in the model is the ninth variable, so to speak, that contributed to the risk score.

What we observed then is, in computing the risk score, there was a range of scores that would be between 0 and 15, so 0 being the lowest risk possible, of course, and 15 being the highest. But what we observed is that in the trial as such, no one had a risk score above 11, so the range turned out to be 0 to 11. And what we observed was that patients who had a low risk, that is to say their point score total or the COURAGE Risk Score total was 0 to 3, comprise 30% of the patients in COURAGE; and they had an observed 9.3% death MI risk over 5 years and that annualizes to just about 1.9%—call at 2%—per year. The intermediate risk group had a COURAGE Risk Score of 4 to 6, and this comprised 60%—or actually to be perfectly specific 59% had an intermediate risk score between 4 and 6. And in this group, the 5-year death and MI rate was 18%, so you know, that was about 3.5% per year. And then there was a smaller subset—11% of patients—who had a high courage risk score of 7 to 11. And this subset had the highest death MI risk over the 5 years of follow up. So they had a risk for death MI of 36% or about 9% per year. So you can see that the low-risk group was about 30%, the intermediate group 60%, the high-risk group 10%. And there was a 4-fold differential in terms of death of MI risk over the course of the 5 years of follow up. 

So why is this important? Well, I think this is important because we recognize that from the most recent ISCHEMIA trial findings, we could find no incremental benefit of the early invasive strategy of revascularization plus optimal medical therapy vs optimal medical therapy alone in a large 5179-patient trial. But by the same token, the question arises, do you have to treat everybody the same way? And I think the answer to that is no, not necessarily. I think if you have a really low risk subset of patients, then you probably would be, perhaps, inclined to treat them with a statin, treat them with aspirin or an antiplatelet inhibitor, and then maybe you know you would not need to use other forms of secondary prevention in a group of patients who only has a 1.9% risk of death or MI over 5 years. Conversely, in the high-risk group, this is the group that you'd really want to double down on medical therapy. And by that I mean we would give a statin, ezetimibe, maybe a PCSK9 inhibitor. In those with low HDL, of course, we'd be interested in trying to raise the HDL, or perhaps using a new medication like icosapent ethyl, which in the REDUCE-IT  trial recently showed that there was a significant event reduction and patients with high triglycerides and many patients with low HDL have high triglycerides.

So my point here is there is a graded scale of risk in patients with stable ischemic heart disease that this tool will hopefully help clinicians to better judge which subset of patients warrant the most intensive therapy. And I might add that in a very high risk stable ischemic heart disease patient—even though COURAGE and ISCHEMIA did not necessarily show an event reduction with revascularization certainly—I would be inclined to think about that in this high-risk group because clearly having such a high rate of events almost 10% per year is really unacceptably high in my opinion. So I think, in that context, I would probably be more inclined to consider revascularization, particularly since we know from the ISCHEMIA trial that spontaneous MI was reduced late during follow up after the first year or two in the trial. And perhaps, that would be a justification in the high-risk subset to consider revascularization to further reduce the likelihood of myocardial infarction.

In any event, thank you for allowing me to review the principal findings for the COURAGE trial. Again, just to refer our readers, this is now published online in the American Journal of Cardiology. I think you’ll find it a very interesting article to read, and I hope that our listeners and viewers will make an opportunity to try to see if this rescore may help them in the care and management of their patients. Thanks very much, Bill Boden signing out.

Submit Feedback