Activity Transcript

Charles P Vega, MD: Hello, I'm Charles Vega, and I'm a health sciences clinical professor of family medicine at the University of California at Irvine in Irvine, California. Welcome to this Medscape program, titled "Patient Selection for Aspirin Therapy: Implications for Primary Care." Joining me today is Dr Michael Blaha, who's a Professor of Medicine and Director of Clinical Research at the Ciccarone Center for the Prevention of Cardiovascular Disease at Johns Hopkins University in Baltimore, Maryland. Welcome, Mike.

Michael Blaha, MD, MPH: Thanks. My pleasure to be here today.

Dr Vega: Great. So we know, and I certainly don't have to tell you, Mike, that aspirin is a very well-established tool that we use to prevent further events in patients who have already had a cardiovascular event. Maybe it's a revascularization, a stroke, a myocardial infarction (MI). Secondary prevention with antiplatelet therapy, such as aspirin, is really important. It prevents further events and can also prevent mortality among these patients. So really, we see the vast majority of patients for secondary prevention should be receiving a antiplatelet agent such as aspirin.

What about an aspirin in primary therapy? There the data are not as strong. It's more mixed in terms of clinical trials. Honestly, 2018 was a really rough year for aspirin as primary therapy among individuals who may have been at elevated risk for a cardiovascular (CV) event, but didn't have a previous CV event.

I'm just going to cover a few trials as context for our conversation. First, there was the ASCEND trial. ASCEND, to me, was the trial of these 3 that really had the most positive results for aspirin therapy. It included approximately 15,500 patients with diabetes, but no history of CV disease (CVD). This study was a randomized trial, so patients received either aspirin or a placebo, and aspirin was associated with a reduction of approximately 12% in the rate of serious vascular events. That's MI, strokes, transient ischemic attack, and death from vascular causes.

So for a composite of those, aspirin was moderately effective, but it was also associated with a higher risk of major bleeding of 29% overall compared with placebo. A lot of that was gastrointestinal (GI) bleeding, and so practically for me, because I see a lot of diabetes in my practice, it relates to my shared decision-making approaches. Which one is more important to the patient? Is it the potential risk of a stroke and MI, or is the risk of GI bleeding more important, or a major bleeding event? Certainly in patients who have a high risk of bleeding, such as a previous major hemorrhage, I'm not recommending aspirin for primary prevention. So that's off the table.

Then there's the ARRIVE trial. Now this had over 12,000 patients, and they were judged to have a moderate CV risk at baseline; between 10% and 20% is what the trialists were shooting for. Here they found no significant difference between aspirin and placebo in the rate of CV events.

Overall, the rate of events was lower in this cohort than they thought it was going to be, and that may be just the more modern approach to management of patients with moderate risk of CVD – more statin use, more lifestyle interventions. So that might have played a factor to under power their study to show a difference between aspirin and placebo in that CV outcome. But again, aspirin did increase the risk of GI bleeding events vs placebo. So you had little benefit, but another risk.

Then ASPREE was another very interesting study because here they looked at people who were older, patients over 70 years of age, or over 65 if they were black or Hispanic. Again it was randomized to aspirin or placebo. Aspirin didn't have a benefit vs placebo in terms of overall mortality, rates of dementia, rates of disability. There was a non-significant reduction only in the risk of CVD or CV death. But again, aspirin was associated with a higher risk of a major hemorrhage. In this study, I think the surprise finding was that there was a higher risk in the aspirin vs placebo group in terms of overall mortality. What was even more surprising to me is that was really fueled by a higher risk of cancer-related death. As far as I know, I don't think that has been fully explained or elucidated as to how that association happened, but it really makes us think that we probably don't want to use aspirin among those individuals at age 70 and older as primary prevention based on the largely negative results of this trial.

A meta-analysis was completed in 2019, after these trials, which bears out what we just described. Really, it's a trade-off. If you're thinking about using aspirin as primary prevention, there probably is some benefit overall across randomized trials in terms of a composite CV outcome, particularly related to the individual events of stroke or MI; less so in terms of mortality, either all-cause mortality or CV mortality. Aspirin does not seem to be beneficial in those randomized trials. The cost is a higher risk of major bleeding, and that includes GI bleeding, which is more common, but also includes a higher risk of intracranial bleeding. To me this is a lot more serious, although also rarer, but it is something that certainly goes into the calculus of decision making with our patients. Aspirin is not a completely benign therapy, and therefore it has to be taken seriously in terms of the cost and benefit ratio, just like any other drug or decision we make.

So Mike, with that background and that context, how do these trials and the overall data contribute to current guidelines for the use of aspirin in primary prevention of CVD?

Dr Blaha: That was so well summarized. Those 3 trials really came out in time for the 2019 American College of Cardiology (ACC)/American Heart Association (AHA) prevention guidelines, to incorporate new data into the recommendations, and of course, the recommendation for aspirin and primary prevention changed. Now it says that low dose aspirin might be considered for the primary prevention of atherosclerotic cardiovascular disease (ASCVD) among selected patients between the ages of 40 and 70 who have a higher ASCVD risk but not an increased bleeding risk. The recommendation for using aspirin was downgraded from a IIa recommendation in prior guidelines to a IIb recommendation, which is kind of a weak, "consider" recommendation in this new guideline. What's a little less clear is how to define that high ASCVD risk group, and even how to define the increased bleeding group. But that's where you pointed out that shared decision-making and other tools are needed to drive this decision.

How do we identify these patients? The good thing is that the guidelines also point to innovations in risk assessment. The cholesterol guidelines, for example, incorporate the idea of the pooled cohort equations (PCE), the 10-year ASCVD risk. And also risk-enhancing factors, things that aren't in the PCE but drive risk, for example, at high lipoprotein (a) (Lp(a)) value, but importantly, the coronary artery calcium (CAC) score, as ways to inform the clinician patient discussion and make decisions in primary prevention.

So for example, the cholesterol guidelines say if you're between 5% and 20% risk, that's a broad borderline to intermediate risk group, particularly if you have a risk-enhancing factor, let's say South Asian ancestry. Consider a CAC score to help further delineate risk. If the score is 0, you can be conservative in your management, probably focusing on lifestyle. If the score is higher, particularly > 100, treat with a statin, a moderate to high intensity statin, to mitigate that risk. So that's where we stand with the guidelines right now and how the CAC score works into shared decision making to figure out risk and benefit. That's the summary of the guidelines, and that's where this discussion will go kind of from within this guideline framework.

Dr Vega: I appreciate that review. Obviously, as you said, there's still room to grow in terms of finding the right patients and applying these important data and guidelines to clinical practice. But you are the perfect guy to talk to in this area because your group conducted a recent study to assess the potential value of using that CAC score for guiding allocation of aspirin therapy. So can you tell us about CAC scores in general and then what you found in your study?

Dr Blaha: The CAC score has become the preferred shared decision-making tool for these intermediate risk patients in whom you're uncertain about treatment. That includes statins, as we discussed, and aspirin. So what is the CAC score? It is a non-contrast cardiac-gated computed tomography (CT) that can be done very quickly and reliably in the outpatient setting. It gives you an assessment of the total burden of plaque in the coronary arteries. The score is directly correlated scored as to the burden of plaque. The higher the score, the more plaque you have, and therefore the higher risk you have. That's how we use the test. How much disease do you have, and therefore how much risk do you have? It's a great predictor of ASCVD events independent of things like the PCE or other risk enhancing factors.

We studied patients who were in the Multi-Ethnic Study of Atherosclerosis (MESA). This is a well known NHLBI-funded 7,000 or so patient prospective cohort study which enrolled community dwelling, asymptomatic adults into a long-term follow up for ASCVD events. All patients at baseline in MESA received a CAC score. These people were 4 ethnicities, non-Hispanic whites, blacks, Chinese, and Hispanic, aged 45 to 84 years. They were followed in this case for 14 years after the baseline assessment, including the CAC score, for ASCVD events. Of course, all the patients got a 10-year risk assessment at baseline using the PCEs. We were able to separate people into the low risk group, the borderline to intermediate risk group, and the high-risk group. Then what I'm going to do is show you the data on the interplay between the PCE and the CAC score for prediction of risk and then model who will benefit from an aspirin in primary prevention.

The groups that we're looking for are CAC scores of 0, well known to be predictive of a low risk state, scores of 1 to 99 that are mildly elevated, and also scores above 100 or 400, which are considered moderately to severely elevated calcium scores. We're following up for similar composite events as you described in the clinical trials. So we're making an apples to apples comparison.

Importantly, we also assessed bleeding events that follow up in MESA, major bleeding events, using International Classification of Diseases (ICD) codes for hospitalizations for major bleeding. We were able to assess both the benefit side with ASCVD, the events, as well as the risk side, the major bleeding events, and then start to model who might benefit from aspirin in primary prevention.

The analysis was to look at patients and their follow up for events and bleeding, and apply from the meta-analysis that you described the risk reduction that we see with aspirin as well as the risk increase that we see in terms of bleeding. And we balanced those things and calculated a number needed to treat (NNT) to benefit from aspirin to prevent an atherosclerotic coronary disease event over 5 years, or a number needed to harm (NNH) to cause 1 bleeding event over the same timeframe, and then start to balance those.

For this particular study, we looked at 3500 or so patients who met the criteria for the guidelines, who were aspirin naive, between the ages of 45 and 70, and not at high bleeding risk, as in there was no history of bleeding, GI bleeding, cancer, et cetera. Of those, only about 316 patients would have qualified for aspirin under the new guidelines as being in a very high-risk state. So a fraction of those would have been identified as an aspirin candidate based on the PCE.

So here's a representation of the interplay between the PCEs and the CAC score in our study. You can see that in the patients who had a ASCVD risk score of above 20%, a large fraction of those have CAC scores of 0 and were actually lower risk patients. Also, patients who were in the lower end of the PCE spectrum, many of those patients had higher calcium scores, hinting at that additive benefit of knowing the CAC score in the risk discussion.

Here are the primary results. What is along the bottom of the slide that we're showing are the PCEs. On the left side, in the NNT and the NNH from giving a baby aspirin, or low dose aspirin, to these patients and following them over time in this study.

First of all, I want to show you the blue bar, which is all the patients in the study, and then the green, yellow, and orange bars, which represent the different CAC scores. What you can see in the blue bars is that whether the patients were low risk, intermediate risk, or high risk, the PCEs were not able to identify a patient group that got a clear net benefit from aspirin. The NNH was lower in some cases or equal to the NNT, creating a situation where there's no net benefit from aspirin.

Because bleeding goes up with risk, as do ASCVD events, you can't find a group that reliably benefits from aspirin based on the PCEs, including the 3 trials that you mentioned. However, when you include the CAC score into the equation, it predicts CV risk, but not bleeding, you start to identify groups that have a clear net benefit from therapy who have an NNT that's lower than the NNH. That was whether you're in the low risk, the intermediate risk, or the high-risk group. With patients who have CAC scores >100, and certainly those with a score > 400, you'd have an ASCVD risk that's sufficient to make the NNT to be lower than the NNH, creating a net benefit situation.

This is starting to eke out who we think will get a reliable net benefit from aspirin for primary prevention Very importantly, if you have a CAC score of 0, clearly the NNH is lower than the NNT. We're probably doing a disservice to our patients who have the lowest calcium scores, like a score of 0. You are probably more likely to harm that patient with aspirin than you are to benefit.

So to summarize, I think we can say that the CAC score can identify patients who are more or less likely to benefit from therapies in primary prevention such as aspirin. I think we'll see them used more in the future and incorporated into future guidelines as the shared decision-making tool.

What I'm really excited about is the clinical application of these results which I think we can move into clinical practice right away. In fact, I already use CAC scoring in my practice to find the patients who are more likely to benefit from aspirin for primary prevention and, importantly, the patients less likely to benefit who might already be taking aspirin for another reason, but may not be getting a net benefit.

Chuck, I'm curious from a primary care perspective, what do you think about translating these results into practice? Because you're always looking for more tools to make that shared decision making with your patients.

Dr Vega: That's a great point. Thanks for sharing that. It's very powerful data, and I can see the applicability. I think in some primary care offices, patients are coming in and saying, "I want this coronary artery testing." That's not my practice. I have some experience with people who have undergone CAC testing, and I do see the value for those patients where you're really on the fence. I find it's a lot easier in my practice to wean people who are aged 78 with minimal CV risk factors from taking an aspirin, or 35 years old, want to be super healthy, their father takes aspirin a day because they've had 2 stents placed in their coronaries, so they're taking aspirin. Those are the low lying fruit that I see. But I have patients who have questions. They have questions about the cost, about insurance coverage, about the ease of use. You mentioned some non-contrast studies, so I think that's a major important step in not having to go to the lab beforehand and coordinate that level of care, and it's relatively rapid. But you know, how has it been accepted in your practice and in more from a cardiology perspective?

Dr Blaha: I find that patients actually accept it quite a lot. They commonly come to me wanting to know their CV risk, whether it's the family history or another condition. They're very curious about their risk. The CAC score, like you said, it's easy to do, which is important. It's a non-contrast CT scan. It can be done on any modern multi-detector CT anywhere in the outpatient setting. It takes me about 10 to 15 minutes of total room time, and only a second of acquisition time. All the patient needs is electrodes on their chest. They can, commonly in my practice, get it done on the way out of the clinic. They leave my clinic, they go get the test done, I call them that night with their results. So they seem to really like that. It gives them rapid feedback and some quick decision making.

Now, one of the problems is that it's not commonly covered by insurance in the United States. It's covered by some private insurers, but not all. Part of the reason for that is it's such an inexpensive test. We charge $75 for all patients for the test, regardless of insurance coverage, just to make it accessible to all. That's easy to justify based on how quick it is and how easy it is to interpret. So I think the average, what we see around the country, is probably between $75 and $150 in places that are trying to really use this on a regular basis. But it's worth remembering how easy this test is to do and interpret. You can just get a score back and there's not any higher level of interpretation. The score is either 0 or elevated, or how elevated it is.

Dr Vega: I think that it is something that you're describing that it's relatively easy to interpret. I've found it really interesting that even for those individuals with a calculated risk above 20%, that score of 0 may be a consideration for not using aspirin therapy. Because for those people at higher risk, it's been something that certainly we've preferred to talk about as an automatic, unless there was a clear contraindication. That may save a lot of individuals a bleeding event that they did not need to have, which I think is really important as well.

We're running out of time, so I feel obliged to ask you about other tools that might augment our understanding of patients' CV risk in primary prevention. Things like ankle brachial index (ABI) and extended lipid profiles. Do they still have a place in this assessment, or has the CAC score really supplanted them with evidence that is a superior tool?

Dr Blaha: In the ACC/AHA guidelines, the CAC scores clearly supplanted those other tests. In fact, there's a class III recommendation for things like carotid intima-media thickness (IMT) testing or advanced lipid testing for everyone in this situation because the CAC score is clearly the better predictor of risk. So I think we're narrowing down. We're not increasing the range of tests we're ordering, we're actually narrowing it down to something that we really think will benefit the patient and hopefully be cheaper than more tests. So things like serum markers and C-reactive protein may be not as helpful. ABI is not as helpful. Routine carotid ultrasound is not as helpful. That's where the guidelines are pointing us right now is if there's a question, the CAC score is the way to go.

Dr Vega: It sounds like we need to use that score more frequently in primary care so we can at least get the conversation started. Of course, we can always refer to cardiology if there's ongoing questions or a second opinion is necessary. But it is really interesting that that CAC score has emerged. Thanks to some of the data you shared with us, I always find that the PCEs get you in the ballpark of your patient's CV risk, but that doesn't get to the level of granularity that you need where you really get insight as to what that individual's risk may be, certainly not every person who walks into the clinic. Therefore I really do see the value in using CAC scores.

Remember, to go back to that ACC/AHA recommendation, that we should be considering aspirin therapy for those individuals between ages 40 and 70, who are at elevated CV risk. We should be thinking about, using the CAC score, just like we used our PCE, to estimate the risk. At the end of the day it comes down to shared decision-making model with your patients. They have to understand the potential benefits and the potential risks of ongoing aspirin therapy.

Mike, thank you very much. This was a great discussion. I really appreciate it.

Dr Blaha: My pleasure, Chuck.

Dr Vega: Thank you to the audience as well for participating in this activity. Hopefully you found it interesting and very practical and applicable to your practice. Please continue on to answer the questions that follow and complete the evaluation.

This transcript has been edited for style and clarity.