Following is a full, unedited transcript of the conversion between Andrew Perry, MD, of Washington University in St. Louis, and Mike Rich, MD, at the same institution. The recording and introduction may be found here.
Dr. Andrew Perry: Thanks for meeting with me today, Dr. Rich. First is the broader question of who should go for stress testing?
Dr. Mike Rich: Stress testing is typically used to try to decide whether a specific patient is likely or unlikely to have coronary artery disease. Often, this is going to be an individual, perhaps, with some risk factors for coronary disease and some symptoms that raise the question of whether or not they have coronary disease.
As a general principle, individuals who have very low likelihood of having coronary disease -- such as a healthy, young person with a paucity of risk factors and obviously atypical chest pain -- should not have stress testing. Conversely, an individual who has a very high likelihood of having coronary disease such as an older person, typically an older male with multiple risk factors and classic exertional angina has a very high likelihood of having coronary disease, and therefore, stress testing to decide whether they have coronary disease or not is typically not going to be helpful. The reason is that even if the stress test is negative, it's more likely to be a false negative rather than a true negative.
Now as we will discuss later on, that type of patient might still benefit from stress testing, but it would not be primarily to determine whether they have coronary disease or not, but rather to determine how severe their coronary disease is and whether or not they might benefit from additional testing such as coronary angiography.
Dr. Andrew Perry: Perfect. Let's spend a minute and talk about that, how we estimate someone's pretest probability. The ACC/AHA guidelines has a table for us that based on age ranges for men and women, and it divides them up into, say, a man who's been the age of 50, 59, whether he has typical angina pectoris or atypical angina, and then it gives you a low, intermediate, or high probability. How do you define, per se, a typical angina versus an atypical angina versus non-anginal chest pain?
Dr. Mike Rich: Right. To follow up on my previous comments, the ideal person for a stress test is somebody who has a pre-test that is, before you do the stress test, likelihood of having coronary disease that is somewhere in the range of, perhaps, 20% to 80%. So what is that population?
As Dr. Perry was suggesting, we have some tools to try to provide some insight into the likelihood that a given individual has coronary disease, and we look at age and gender. The older a person is, the more likely they are to have coronary disease. Men, in general, are more likely to have coronary disease than women. Then we also look at the nature of the symptoms.
Classical angina where somebody has exertional chest discomfort, pressure, tightness, heaviness that come on consistently with exertion at some level. A person might say, "Well, every time I walk up this hill near my house, I get this discomfort in my chest, and it resolves when I rest for a couple of minutes, and then if I do the same thing again the next day, I'll have the same symptoms." That's classic angina and that is consistent with coronary disease.
Another type of symptom might be some chest discomfort that has some features that are typical of coronary disease, but others that are not. For example, the symptoms may not come on. Only with exertion, the person might have similar symptoms at rest or at night or they might not be described as a pressure or a heaviness. It might be some other description of it. It may be somewhat of a burning quality, but can't really be sure that it's not coming from the heart.
Then there's a third type of chest pain, which is clearly atypical and unlikely to be coronary disease, and this would be chest pain that is sharp in nature, focal, one particular spot in the chest, it's reproducible by palpating that area, and often will be pleuritic. The person takes a deep breath and that aggravates the pain. That type of pain is very unlikely to be due to coronary ischemia.
Then the fourth category would be the person who has no symptoms whatsoever. They don't describe any chest discomfort or other symptoms that might be considered an angina equivalent. We take the nature of the person's symptoms in combination with their age and gender and try to come up with a pretest likelihood that the person has the disease.
Just to give an example. A 50% pretest likelihood which would be an ideal candidate for considering a stress test might be a middle-aged male with a couple of risk factors and symptoms that are possibly anginal in nature. Or alternatively, it might be an older woman with fairly typical angina also in the context of a couple of risk factors. Those would be individuals who would likely have a 50-50 chance of having coronary disease and whom, therefore, might benefit from stress testing.
Dr. Andrew Perry: Then additionally as kind of a summary, what we've been kind of discussing without really outright saying it is about Bayes' theorem and using your pretest probability to then guide your testing. Really, the prevalence of your disease has a big impact on your positive predicted value. So could you just kind of comment on how we're using pretest probability in order to select our population of prevalence?
Dr. Mike Rich: Right. Bayes' theorem states that for any less-than-perfect test and stress testing is certainly a less-than-perfect test, that the posttest likelihood of having the disease or not having the disease based on the results of the test is contingent upon the pretest likelihood of having the disease. For example, stress testing with imaging, traditional stress testing, we'll talk more about that in a little bit, might be considered to have roughly 80% to 90% sensitivity and 80% to 90% specificity.
If you take an individual who has a 50% pretest likelihood of having the disease and the test is positive, that will increase their posttest likelihood of having the disease to the 80% to 90% range. Similarly, if the test is negative, that will reduce the likelihood of them having coronary disease, significant coronary disease, to the 10% to 20% range. So you get a nice separation of that 50% pretest likelihood to high likelihood or low likelihood after completing the test.
Conversely, if you take an individual who has a very low pretest likelihood -- such as a young, healthy person with atypical chest pain and a paucity of risk factors -- and you do the stress test, if it's negative, then you've confirmed that they, in fact, have a very low likelihood of having the disease. However, if it's positive, then because of Bayes' theorem and this issue of pretest likelihood and the impact of that on interpretation of the result, if the test is positive, it still is more likely that it's a false positive than a true positive. This, unfortunately, can then precipitate a cascade of additional testing or perhaps initiation of therapies when, in fact, the person doesn't really have the disease. So that's the primary reason why stress testing, in general, should not be performed in individuals who have a quite low pretest likelihood of having the disease.
Dr. Andrew Perry: As the way I look at it, for ordering now, choosing what stress test to use, there are kind of two factors to be thinking about. First is the stress portion, and then two, how I monitor their response. First, I can either stress them by stressing them with exercise or pharmacologically, and then second, how I measure their response, either electrocardiographically or with some type of imaging. I guess can you walk us through maybe your approach to choosing the appropriate stress test?
Dr. Mike Rich: There are several considerations. One is that if the person is able to exercise, one should always do an exercise test. The reason for this is that you get quite a bit more information from an exercise test. First, it's physiologic as opposed to any kind of a pharmacological stress test. You get information about that person's exercise capacity. If the test is positive, you get information about how much exercise it takes before the person starts having symptoms or ECG changes, and there's considerable difference between a stress test that is positive within the first two minutes of exercise as opposed to a person goes 10 or 15 minutes of exercise and has some symptoms.
It also will influence the likelihood that you would pursue additional testing. If somebody has a clearly positive stress test after two minutes, there's a reasonable likelihood that you'll proceed with coronary angiography. Alternatively, if the person goes 10 or 12 minutes, then it would be reasonable to pursue a strategy of medical management as an initial option.
The next question is what type of stress test? The guidelines recommend that in an individual who has a normal ECG and is not taking any medications that might influence the ECG response to stress testing, that the first test that should be performed is just an exercise ECG, an exercise treadmill and without any imaging. The reason for this is that although it's only associated with moderate sensitivity, roughly 70% or so, that if the person has a good exercise tolerance and the test is negative, that no further test is needed at that point.
Alternatively, if the test is positive, then one can decide, depending on how positive and how early into the testing it's positive, whether or not to follow up with an additional stress test with imaging to further define the location and severity of ischemia or, in some cases perhaps, proceed directly to coronary angiography.
The exercise stress test with ECG only, no imaging, has some important advantages. Most significantly it's substantially cheaper than any of the imaging protocol. It typically runs an order of a couple hundred dollars to do that type of a stress test as opposed to a couple thousand dollars for any of the imaging stress tests, or more. That's the first choice.
Beyond that, if either the person has an abnormal ECG at baseline, and therefore, is unlikely to have a reliable stress ECG or if the person has an exercise ECG and it's abnormal, then the next step would be to do a stress test with imaging. The possible types of stress tests is expanding, and now includes stress testing with PET or MRI or even CT, but for purposes of this discussion, I'm going to limit it to the more traditional stress echo or stress nuclear types of studies.
Between those two, there are really four options: an exercise echo, a pharmacologic echo, typically with dobutamine, a stress nuclear study with exercise, or a pharmacological nuclear study, often also referred to as a myocardial profusion imaging or MPI. Of these four options, in experienced centers, they have roughly equal sensitivity and specificity, again, in the 80% to 90% range and predictive accuracy. In terms of the results of the stress or interpretation of the results of the stress, there's little to choose between those four different options.
So how do we decide? I think that first, as I mentioned already, always best to perform an exercise test if the patient is able to exercise. The next question is well, should we do an echo or a nuclear study? I think that in most cases the echo test is preferred and the reasons for this are that, first you get the additional information from the echocardiogram, including information about wall motion at rest, wall thickness, atrial size, valve, any problems with valve function, and so forth. In fact, often individuals who you want to do a stress test, you may also think that you want to get an echocardiogram. To some extent, a stress echo "kills two birds with one stone."
For the reasons I mentioned, in most cases, you'd like to get an echo study, in part, because you get more information from it, but also it does not involve any radiation exposure, so that's a significant advantage. It's somewhat cheaper than nuclear studies, and it's also quicker. A typical stress echo, all of the images and so forth can be completed within no more than an hour, including the posttest observation period. Whereas with nuclear studies, it can take several hours. For an individual, for example, who's working and trying to schedule this stress test during working hours, it's less time off of work. That can be an important advantage from the patient perspective.
Dr. Andrew Perry: We kind of talked about the asymptomatic patient and then patients with symptoms who are this intermediate probability. What about the very, very high-risk patients? Is there anyone that you just say, "No stress test, straight to cath?" What are those patients like?
Dr. Mike Rich: Right. I think that individuals who have high-risk symptoms such as acute coronary syndrome, and this is not somebody who's coming in with stable angina. They're having symptoms that are progressing and occurring at rest, and you have a high enough pretest likelihood that you think that they almost certainly have coronary disease, that in those cases, omitting the stress testing and proceeding straight to another type of imaging would be reasonable.
I'm waffling a little bit here because I think that an alternative to traditional invasive coronary angiography, would be coronary CT angiography, which is significantly less invasive, has high sensitivity, and this is a situation where it might be overly sensitive.
If, however, coronary CT is negative, it has a very high, close to 100% negative predictive value. If the CT angiogram is negative, and the image quality is good, the likelihood, then, an individual has significant coronary disease is quite low and you can, in such situations, avoid proceeding to invasive coronary angiography.
But if that's not a good option in a given case, then yes, I think there are situations in which the risk is high enough that I would proceed straight to coronary angiography, and again, this is going to be significant variability amongst cardiologists of what the threshold might be between going for stress test first or CTA or coronary angiography.
Dr. Andrew Perry: Okay, I think we've had a good discussion so far. I have a couple cases that I want to discuss with you that'll probably highlight some of these issues here in practice. For example, here's a patient that I saw in clinic just a couple of months ago, but, her history goes back a few years, so we'll start there.
It's a 38-year-old female and an immigrant from Bosnia. She was complaining of some squeezing chest pain that was radiating to her left arm. She said when she got this kind of pain it would be associated with a shortness of breath, some dizziness and sweating, kind of variable onset, with activity, sometimes without activity. She is a smoker. She also reports that her sister had a stroke when she was in her 20s, but I don't have any details about that. She initially went to the emergency department where they checked an EKG, which was unremarkable. They checked some troponins, which were negative. In this case, is there a role for stress testing and what would that be?
Dr. Mike Rich: Yes, I think so. We have a relatively young woman who presents with fairly typical chest pain and associated symptoms, shortness of breath and diaphoresis. One atypical feature is that the symptoms sometimes are not associated with exertion, and so it's less of the classic angina, perhaps, in that regard. But she has some risk factors. She's a smoker, which is an important risk factor particularly for women, and some positive family history for atherosclerotic disease, perhaps a stroke in her 20s, which might not have been due to atherosclerosis, but in any case, cardiovascular disease in the family at young age.
We're not told anything about other risk factors such as her lipid profile or blood pressure or diabetes. But even without that information, I think that she probably is in the 20% or higher threshold that we discussed earlier, and I therefore think that doing a stress test would be reasonable. If her EKG is normal, then I think that starting with an exercise treadmill test would be reasonable. If her EKG is, at baseline, not completely normal, then a stress test with imaging would be reasonable.
Dr. Andrew Perry: Gotcha. So there in the emergency department, she went for a stress echocardiogram. She performed 7 METs and it was reported as being negative for ischemia in the comments below. Her Duke Treadmill Score was reported being a 6. She was discharged to home. Can we just take a second and talk about what that means, having exercised 7 METs, and then what is a Duke Treadmill Score?
Dr. Mike Rich: Right, so 7 METs for a 38-year-old is relatively low exercise tolerance. If the person was in their 60s, then 7 METs, which basically means they went seven minutes on the Bruce Protocol on the exercise treadmill. There's a close correlation between the number of minutes that a person exercises and the number of METs, which is the metabolic equivalence of exercise performed.
The Duke Treadmill Score is a simple algorithm for assessing risk based on three factors during the stress test. The first is the number of minutes performed on Bruce Protocol. That's equivalent to the number of METs here. That's 7. Then from that, you subtract five times the maximum ST segment deviation, either depression or elevation in any lead except for AVR. So fives that is subtracted from the METs or the minutes of exercise.
Then additionally, you subtract four times any symptoms that the person has. If they're asymptomatic, that's a 0. If they have some discomfort, but it's not limiting, in other words, that's not the reason why they stopped exercising, then that's 1 point, and if they have symptoms that are fairly typical of angina and are the reason why the person stops exercising, then that's worth 2 points. You'd multiply that by 4 and subtract that from the METs as well. In this particular case, she exercised for 7 METs and we're not told whether there were any EKG changes.
Dr. Andrew Perry: There were no EKG changes.
Dr. Mike Rich: So we're not subtracting anything for that, and we're not told whether she had any chest pain. But we'll assume basically that she didn't, and we're not subtracting anything for that, so that would actually give a DTS, Duke Treadmill Score, of 7. Here it says 6, so same ballpark.
Then interpreting that number, the higher the score, the better, and generally a score of 5 or higher is considered low risk. A score of -10 to +4 is considered intermediate risk, and a score of -11 or lower is considered high risk for having significant or severe coronary disease.
Dr. Andrew Perry: Okay. That was about six years ago. She goes home and over the next six years she continues to have this chest pain intermittently off and on. She presented to me two months ago with now a more recent, one-month history of the chest pain just becoming much more worse. She was starting to have it more at rest and just being more severe in intensity as well.
She went to the ER, actually, and they put her over in observation, but she didn't want to wait for the stress test to be done in the morning, so she went home. Noted on her EKG taken there, though, were noted some new Q waves that had developed in V1 through V3. From this point, what would be your decision point about whether go for stress testing or maybe would you even be more aggressive for her at this point?
Dr. Mike Rich: I think that we have a variety of options here. I think that the new Q waves are concerning since they suggest that she may have had a interim myocardial infarction, anteroseptal myocardial infarction. She's having worsening symptoms, which if the nature of the symptoms is similar to what she reported previously, they can be interpreted at least as being consistent with exertional angina.
I think that one can take three possible approaches. First would be to do another stress test. We have previous stress tests, perhaps, for comparison, so that might be helpful. Second would be to go straight to a coronary CT angiogram, which would provide some information about her likelihood of having severe coronary disease. The third option, which I'm sure that some cardiologists would support in this particular situation, and which I think would be reasonable also, would be just to go straight to catheterization given the severity of her symptoms, the progression, and the new ECG findings just to define her coronary anatomy and make management decisions based upon that. I think any of those options would be reasonable.
Dr. Andrew Perry: Okay. So she came to me in the urgent care clinic a couple days later, and I saw her and I set her up for a stress echocardiogram, which she went and performed and she performed 10 METs on the echo. I got a call from the cardiologist saying that it was markedly positive for ischemia in the LAD territory and recommended that I pursue it and follow it up with a cath.
She gets admitted to the hospital for a cardiac cath, which demonstrates mild disease in all three vessels. 20% lesions in the LAD, in the circumflex, and the RCA, which was a little unexpected for me. Maybe you can comment on how the stress testing leading to this result, and then what the cath shows and kind of how this all maybe fits in together.
Dr. Mike Rich: I'd like to just return to the decision point about how to manage the patient after she has the positive stress test. She goes for 10 METs and has what's described as a markedly positive test for ischemia in the LAD distribution. I think that we have three options. One would be simply to initiate medical therapy if she's not already on beta-blocker, make sure that her risk factors are under good control, as we discussed previously, and see how her symptoms respond to more aggressive medical management.
The second option would, again, be to do a coronary CTA or, again, the option of going straight to coronary angiography. I think that any of these three approaches would be reasonable. In this case, the decision was made most likely due to the fact that the test was read as markedly positive in the LAD distribution. This is worrisome, and she has additionally the new Q waves in the same distribution. There's relatively high concern that she may have very severe disease and that an intervention might be appropriate. The decision is made to take her to the cath lab where, in fact, she is found to have no significant epicardial coronary disease.
How do we interpret that finding? First of all, does this mean that the stress test was a false positive? I would say the answer is not necessarily, particularly in a woman. She could have small vessel disease, which is giving both symptoms and an abnormal stress test in the absence of significant obstructive epicardial coronary disease. In which case, it's not really a false positive.
It's a false positive in the sense that she didn't have obstructive coronary disease, but her symptoms are still truly ischemic and will likely respond to conventional medication management beta-blockers, most importantly. Beyond that, though, there is a possibility that her symptoms aren't ischemic at all, and if that's the case, then it is a false positive test. As we discussed earlier, the specificity of stress testing is in the 80% to 90% range, and what that means is that there's a 10% to 20% likelihood of a false positive. It's not trivial. We see false positives not uncommonly, and again, that's part of the reason why we use Bayes' theorem and the pretest probability of having disease in deciding whether or not to do a stress test in the first place.
Dr. Andrew Perry: Just to kind of circle back, I pulled up the report from her stress echo that I had sent her for. On the EKG, there were no ischemic changes noted on there. So her Duke Treadmill Score would probably be right around 6 of where it was before.
Dr. Mike Rich: Right.
Dr. Andrew Perry: As you said, maybe some medical management would be reasonable at that time given that she's in that low-risk category based on there.
Dr. Mike Rich: Right.
Dr. Andrew Perry: Okay. Then finally, one other case. It's a little different, but it highlights another utility for stress testing that we often encounter here on the wards. This is a 64-year-old male who was admitted to the hospital. He has a history of coronary artery bypass graft in 2015. He's presenting with unstable angina. Chest pain at rest, negative biomarkers of troponins, and his EKG was stable as compared to his prior ones.
He recently just even had a repeat cath. Two months prior to admission, it showed that his LIMA grafted to LAD was patent. He had an occluded superior vein graft of the OM branch and a total occlusion of the RCA distal to the RV1 branch, which had been chronically there in the past. For this patient, known coronary disease, status post bypass grafting. What is the utility of a stress test for this guy?
Dr. Mike Rich: This falls into a different indication for performing a stress test. We're no longer doing the stress test to rule out coronary disease. This person has known coronary disease, and their pretest likelihood of having coronary disease is 100%. They have known coronary disease.
The rationale for doing a stress test, in this particular situation, is to assess the severity of the disease and how limiting it is. In other words, we're interested in how much exercise does it take before the person develops symptoms and how severe is the ischemia that they have in response to exercise? That is often an important piece of information in deciding whether or not to pursue more aggressive medical therapy for management of the patient's disease versus proceeding with repeat coronary angiography.
For example, if the person gets on a treadmill and goes seven or eight minutes and has mild symptoms in a mild to moderately positive stress test, then it would be very reasonable to increase medical therapy to try to control the person's symptoms. If, on the other hand, the person goes two minutes, has three or four millimeters of ST segment depression and marked chest discomfort, then that would be somebody that it would be reasonable to proceed with coronary angiography.
In this particular case, I think that the fact that the person had a cath two months previously, which did not show any revascularizable myocardium, would at least, in my view, make it less likely that a stress test is going to be helpful. I think that if the person is not on maximum medical therapy that that would be the first thing that I would recommend in this particular situation.
If the person is on maximum medical therapy, then the question will be how are we going to interpret the results of a stress test? If it's positive, are we going to do another cath? How are we going to use the information? I think that it would be a difficult decision given that there was a cath two months previously that failed to show revascularizable coronary disease.
Dr. Andrew Perry: Okay. Would your approach change if there were some mild biomarker elevations or some non-specific ST or T wave changes for him?
Dr. Mike Rich: Not necessarily. The approach might change if you didn't tell me he had a cath two months ago.
Dr. Andrew Perry: Okay.
Dr. Mike Rich: If we didn't have that cath two months ago, then I think it would be very reasonable to do a stress test in this particular individual. So he's having increasing symptoms that are described as unstable angina with negative biomarkers, stable ECG, and so the question in this case is whether this really ischemia, and if so, how severe ischemia is he having? In that case, I think it would be reasonable to do a stress test as the next step in the evaluation, and if it's markedly positive, proceed with angiography.
The situation is modified if, in fact, the biomarkers, troponins are elevated, and the higher they are, the more likely I am going to recommend proceeding straight to coronary angiography. Similarly, if the ECG shows new changes consistent with ischemia, how severe are they? How extensive are they? Those are factors that are going to be taken into consideration in deciding whether to go straight to angiography or whether to proceed with stress testing.
In this situation, coronary CTA is probably less likely to be helpful, although, in some cases, it might be useful, so that would be an alternative, but it would be lower down on my list, anyway, of options of how to further evaluate the patient.
Dr. Andrew Perry: Well, thank you very much for your time today and discussing stress testing with me.
Dr. Mike Rich: Certainly, my pleasure.
, a resident physician at Barnes-Jewish Hospital and Washington University School of Medicine, in St Louis.