The physical exam findings of aortic stenosis and how they can help in diagnosis and grading of aortic stenosis are the focus of this episode of AP Cardiology with host Andrew Perry, MD, joined by Shreyas Venkataraman, MBBS, of Washington University School of Medicine in St. Louis and its Barnes-Jewish Hospital.
They share two examples of heart murmurs and discuss likelihood ratios and the physical exam finding that is most consistent with severe aortic stenosis.
A transcript of the podcast follows.
Perry: Hey, everyone, Andrew here. Remember how for almost the last year I've been talking about publishing some episodes regarding physical exam and cardiac auscultation? Well, today is the first of those episodes and we're going to be focusing in on aortic stenosis and discussing the historical findings and the physical exam findings that can help improve our exam and bedside grading of aortic stenosis.
I'm joined by an incoming intern at Barnes-Jewish Hospital, Shreyas Venkataraman. Thanks goes out to the folks over at Thinklabs who produce the Thinklabs One stethoscope. This is a digital stethoscope that I have found very useful in my clinical practice and recording a library of heart murmurs and heart sounds. Please note that all patients provided consent for the use of these recordings. I hope you enjoy today's episode. With that, we'll get started.
This is AP Cardiology and this is your host, Andrew Perry. Hey, everyone. It's Andrew here. I'm joined today with a new member of the team for AP Cardiology and I will let Shreyas introduce himself for you all.
Venkataraman: Hey, everyone. This is Shreyas. I'm an incoming intern in the Barnes-Jewish Hospital and the newest member of the AP Cardiology team. Hey, Andrew. How you're doing?
Perry: I'm good and how are you doing, Shreyas?
Venkataraman: I'm doing wonderful.
Perry: Shreyas is teaming up with me. He's helping me out today with starting our series on the physical exam, and so we're going to discuss about aortic stenosis, the physical exam for aortic stenosis, and which parts of the physical exam are most useful in clinching that diagnosis at the bedside. With that, how about we start with our case?
Venkataraman: Today, we have a patient presenting to the clinic, a 72-year-old man with a history of chronic kidney disease, stage III, and diabetes, who presents with dyspnea on exertion and occasional presyncope.
On further questioning, he denies any chest pressure with exertion, orthopnea, and paroxysmal nocturnal dyspnea. Now, his past medical history is significant for type 2 diabetes, for which he's on insulin, and his family history is not significant. His social history is significant for a 20-pack-year smoking history and he has one to two beers a week. I think we can stop right here.
Perry: Perfect. A 72-year-old man, this is obviously an episode on aortic stenosis, so we're going be biased approaching it. But dyspnea on exertion is a very common symptom and there is a whole differential there, and I would refer you to like the for a better discussion on how to approach dyspnea on exertion.
But when we're thinking about aortic stenosis, there are specific parts of the history that can be helpful -- and the physical exam -- in helping to increase the probability of that diagnosis. Shreyas, through our discussion today, we're going to be talking a lot about likelihood ratios. Do you want to just give us a little primer on how to use likelihood ratios when we're thinking about these points?
Venkataraman: For sure. In Dr. McGee's "," he explains the 2-5-10 rule, which is a great shortcut for likelihood ratios where we can translate pre-test probabilities into post-test probabilities without the use of odds.
The 2-5-10 rule essentially means that a likelihood ratio of 2 increases the post-test probability by a factor of 15, or 15 percentage points. A likelihood ratio of 5 increases the post-test probability by 30 percentage points, and a likelihood ratio of 10 markedly increases the post-test probability by 45%.
If there were 6 numbers, we have to remember they would be 2, 5, 10, 15, 30, and 45. This is for a positive likelihood ratio and an increase in the post-test probability. In contrast, the reciprocal of 2, 5, and 10 -- that is 0.5, 0.2, and 0.1 -- conveniently decreased the post-test probability by 15%, 30%, and 45%, respectively. To briefly summarize, let's remember the numbers 2, 5, and 10 increasing and decreasing the probability by 15%, 30%, and 45%, respectively.
Perry: Beautiful. Love it, and thank you for that review, because in the literature on physical diagnoses the most common thing that's reported is going to be likelihood ratios. It's hard to then think about how that then translates at the bedside.
When we're thinking about this patient that we just discussed -- we have a 72-year-old, he's male, he's diabetic -- that in and of itself is going to increase our pre-test probability for aortic stenosis. All comers in that age range are going to have somewhere between a 5% and a 10% prevalence of aortic stenosis. Really, the historical findings that we present -- dyspnea on exertion, maybe some occasional presyncope -- most of these aren't very helpful.
In fact, they have a lot of information in terms of prognosis. But in terms of diagnosis, they're not super-helpful. Like effort syncope is the only well-studied historical trait or characteristic, and that really only increases the probability of aortic stenosis with a likelihood ratio of 1.3, so that's less than 2. That's less than a 10% increase from our pre- to our post-test probability.
Everything else -- we're talking about the chest discomfort, dyspnea on exertion -- as I said earlier, they're so common that they're really not helpful in differentiating between other cardiac or respiratory causes.
Venkataraman: It's interesting that you would say that, Andrew, because we hold onto these historical findings. But having said that, they do point us to a more accurate physical exam, don't they?
Perry: Yeah. No, they do and so it should tune your mind in and think, "Okay, I need to pay attention to specific things on my exam."
Venkataraman: Let's continue with this case. On his vital exams, his blood pressure's measured at 110 over 60, his heart rate was 90 beats per minute with a regular rate and rhythm, his respiratory rate was 18 breaths per minute, and he was satting at 97%. What do you think about that, Andrew?
Perry: To me, those are pretty unremarkable vital signs and just indicates to me the person has normal blood pressure, kind of a higher heart rate, a little tachycardic. But otherwise, the person is not in shock and they are currently stable, like someone you'd probably see in your clinic.
Venkataraman: What are your thoughts on pulses with patients having a history suggestive of aortic stenosis? Do you think of certain things while you're examining these patients' pulses?
Perry: In evaluating patients with aortic stenosis, I always evaluate their carotid pulses and I'm looking for a couple of things. I'm looking for a diminished pulse, less volume, and then I'm also looking for a delayed upstroke in that pulse. In particular, the decreased volume in the carotid pulse, it will increase your likelihood ratio of aortic stenosis by 2.3, which still is small. It's like a 10% increase in your post-test probability.
Venkataraman: It's interesting that you would say that, Andrew, but another finding that could help us is a delay in pulse in relation to two pulses that can be felt within the body. Two findings that are commonly used are the apical-carotid delay and the brachioradial delay, where the apical impulse is compared temporally with the timing of the carotid impulse for the apical carotid delay, and the timing of the brachial pulse is timed with the radial pulse. Any delay is abnormal. Even the slightest delay that can be felt is abnormal.
This finding is helpful in the way that if there is an absence of an apical-carotid or a brachioradial delay, it markedly reduces the likelihood ratio of aortic stenosis, so it has a likelihood ratio of 0.04 and 0.05, which is amazing. It reduces the probability by around 50%, which is amazing. But interestingly enough, the presence of these two findings doesn't help us much.
Perry: Gotcha. That's a useful test to say that they don't have aortic stenosis and to say that there's no delay in their pulses. They have a normal time in between the brachial and the radial pulse.
Venkataraman: Correct.
Perry: Perfect. Let's go on more with our exam.
Venkataraman: Perfect. On general appearance, he is alert and oriented times three. There is no neck vein distention on head and ENT examination. His lungs are clear to auscultation bilaterally. While we examine his cardiovascular system, on palpation, interestingly enough I feel a sustained apical impulse in the fifth intercostal space at the mid-clavicular line. I don't know what that really means. What are your thoughts about that, Andrew?
Perry: A sustained apical impulse. You're describing the fifth intercostal space in the mid-clavicular line. I'm pretty simple and I think, like I said, near the nipples and not near the nipple. That's pretty close to the nipple from what you were describing, so that's a normal point of maximal impulse.
A sustained apical impulse would indicate that there's increased pressure and it's more delayed. That can happen when the pressure inside your LV, while it's contracting against a fixed gradient, can be sustained for longer periods of time. That might be what you could be feeling when you're examining those patients.
At the same token, with all the other exam findings we've been talking about so far, a sustained apical impulse is also not very helpful. Again, a likelihood ratio of 4, and so that's somewhere around a 20% to 30% increase in the post-test probability of aortic stenosis.
Venkataraman: Interesting. Sometimes you hear a thrill too, don't you? Sorry, if we feel for a thrill, rather?
Perry: Yeah. Another part of the exam is that I can put my hand over the patient's chest, over their sternum. If I feel a thrill, that's indication that there is a lot of turbulent flow in there and more turbulent flow is associated with a greater severity of aortic stenosis.
Venkataraman: We went over a lot of numbers, a lot of likelihood ratios and a lot of percentages. I think this would be a good time to just pause, take a step back and briefly review what we talked about right now. What do you think, Andrew?
Perry: Yeah, let's do it.
Venkataraman: Wonderful. There's no definitive historical finding that increases the probability of aortic stenosis. But having said that, it does help us perform a more directed and more focused physical exam. That's the first take-home point.
Secondly, the absence of an apical-carotid or a brachioradial impulse markedly reduces the likelihood of aortic stenosis -- the absence, that is -- but the presence, interestingly enough, does not help as much. Finally, a sustained apical impulse increases the probability of aortic stenosis by around 30%.
Perry: Beautiful. Now, let's turn towards auscultating. I have two recordings to play for you and we'll be able to compare both of them. Here is the first recording of heart sounds that you might hear when you examine this patient.
Now, here is a second recording of something that you might hear.
Venkataraman: Now, let's play that again simultaneously so that we can compare the two different murmurs heard in two different patients.
Let me try and describe the murmur for you, Andrew. What I heard was -- at least with the more obvious murmur we heard initially -- I heard a harsh systolic murmur. I heard it over the left lower sternal border and over the base of the heart on the right side. What are the things I'd think about when I'm auscultating the heart?
I look for heart sounds. I look for the S1 sound and then I look for S2, and I look for any murmurs that could potentially exist. I take the time to go to a quiet room and listen for murmurs. When I do hear a murmur, I make sure I focus on where I hear the murmur best on the chest.
Secondly, I make sure I understand when I hear the murmur best. I try and understand where the murmur radiates and finally I try to describe the murmur like I had done earlier, where I define it as harsh, and so on, and so forth. What are your thoughts about what we just heard right now, Andrew?
Perry: First-off, the fact that we're examining the patient and that we're hearing a murmur at least should give us a consideration that there is aortic stenosis or that they have some valvular disease. When I examine patients, I feel their pulse so that I know with the timing if the murmur is systolic or if it's diastolic. Here, with this murmur -- you just have to take our word for it that this is a systolic murmur -- and we can hear a crescendo murmur here.
Another thing that we've identified systolic, we've identified that it's harsh. The harshness really is just something that comes with experience of listening to various people with different murmurs about how severe that harshness is.
Then location, as we've all learned in medical school, can give us an indication of where, that certain murmurs for different valves can occur in different places there. But just even the fact that we're hearing a murmur should suggest aortic stenosis. If there's no murmur, that really reduces the likelihood of having aortic stenosis, with a likelihood ratio of 0.1. Shreyas, how do you grade murmurs?
Venkataraman: They're graded on the Levine's scale. He defined it as Grade 1 to Grade 6, with Grade 1 being a murmur maybe only Braunwald himself could have heard. From then on, I think we can understand it as us hearing it better, where maybe Grade 1 is a murmur maybe an attending cardiologist hears best, a Grade 2 being a murmur a cardiology fellow hears best, a Grade 3 murmur being a murmur that a resident hears best, and a Grade 6 murmur being something I can hear best.
Essentially, the threshold to split the Levine scale is a point between Grade 3 and 4, where if there's a thrill present it becomes Grade 4. If there's no thrill, it is Grade 3. I think that's a good way to think about these murmurs. I think back in the day when we didn't have echocardiography, I think this was relevant, but it allows us to have a more intelligent conversation about murmurs when we grade them this way, I guess. What do you think about that?
Perry: Yeah. We probably live in an era where we're not going to hear any Grade 6 murmurs at this point. Probably all that we'll be hearing are Grade 3 and Grade 4 at the most severe of our murmurs. Yeah, I totally agree with you. Grade 1 murmurs are going to be murmurs that I really struggle to hear consistently. Like a nice, quiet Grade 1 mitral stenosis murmur is going to be very hard to hear.
Venkataraman: Perfect. What are you thinking about with this murmur, Andrew?
Perry: Another thing we haven't touched on yet is about the S1 and S2 sounds. In the second example that we'd listened to, we don't hear a clear S2 sound and the absence of S2 is a very specific finding that indicates the severity of aortic stenosis.
If fact, not hearing an S2 increases the likelihood of aortic stenosis with a likelihood ratio of 12.7, meaning an increase in probability of well over 50%, so a very powerful predictor in the severity. In fact, the absence of S2 should indicate not just the presence of aortic stenosis, but should be an indicator that this person likely has severe aortic stenosis.
Venkataraman: From what we've talked about until now, this seems like the strongest finding that helps us direct our attention a particular way. Is that correct?
Perry: Yes, that is correct. The parts that we've hit on most far that people talk about a lot at the bedside are the murmur radiating to the carotids, which we have discussed, isn't as helpful. The harshness of the murmur as well is not very differentiating about the severity of our aortic stenosis. However, the absence of S2 is a very powerful predictor and should be something that we really pay attention to when we're examining patients with aortic stenosis.
Venkataraman: What are your thoughts about the murmur itself?
Perry: The murmur itself, I would describe this as a harsh murmur and I would describe it as a crescendo, so it's getting louder, and kind of a later-peaking murmur that we're hearing. Additionally, you had mentioned that you were hearing this heart sound along the left lower sternal border, the right upper sternal border, and also at the apex.
The fact that we can hear the murmur throughout all these different points where we auscultate the heart commonly is a more powerful predictor of the likelihood of aortic stenosis as well. Whereas if we're only hearing the murmur when we're auscultating over the aortic position, we may just be fooled and tricked into hearing a systolic ejection murmur, which sometimes happens.
Venkataraman: What else do you think about when you listen to this murmur?
Perry: When we talk about radiation of the murmur and it radiates to the carotids, that is something that's also discussed a lot and something that I always examine for as well. However, the utility of that in increasing my likelihood of aortic stenosis is also only mild or moderately helpful. The likelihood ratio of aortic stenosis when the murmur radiates to the carotids is only 2 to 3.5, which, again, is somewhere around a 10% to 20% increase in the post-test probability.
Venkataraman: Yeah. It's not too much. It doesn't really help us. A finding which we usually associate with aortic stenosis doesn't really help us as much.
Perry: In all of these things, like you have a couple of different findings and they can all add up, and so that's when they're helpful. But each one of these findings, in and of itself -- aside from the presence or absence of S2 -- is not very powerful in and of itself, which is why an echocardiogram is usually very helpful in delineating the etiology of the murmur and the severity of aortic stenosis for these patients.
That said, there should be some caveats when we're interpreting an echocardiogram report, because the physical exam findings and the echo reports sometimes may disagree. For example, you may examine someone. You may hear a harsh, late-peaking systolic murmur, radiates to the carotids, and you don't hear an S2.
But when you order your echocardiogram and you're looking at the report afterwards, you might see findings with the mean gradients and peak velocities that are more consistent within the moderate aortic stenosis range and not clearly severe aortic stenosis.
There, it's helpful to realize that there can be some limitations to echocardiogram in the fact that the peak velocities and the mean gradients may be falsely reported as low if the sonographer is unable to get in direct alignment with the flow through the aortic valve, which in some patients may happen. You may not be getting the actual true peak velocity or the actual true mean gradient.
In those cases, if you are having a physical exam that's more consistent with severe aortic stenosis but your patient's echocardiogram is in disagreement with that, that's the time where you need to pause and think, "Perhaps I need more information," and think about a cardiac catheterization study, or possibly repeating the echocardiogram and being more specific in looking for severe aortic stenosis.
Venkataraman: Amazing. We talked quite a bit about auscultation in these patients, so let me briefly review what we talked about. If there's one point you take home, it's the fact that the absence of S1 and, most reliably, the S2, markedly increases the probability of aortic stenosis by as much as 50% to 55%.
There is a very broad pattern of location when we hear for these murmurs. It's defined as an apical basal pattern, where it can be heard on multiple locations in the pericardium. Finally, radiation of the murmur does not distinctly increase the risk of aortic stenosis.
Perry: Beautiful. No, that was a great discussion. I'm really glad that we were able to get together and do this, Shreyas.
Venkataraman: Yep, and I learned quite a bit, Andrew. Thank you for having me here. It was wonderful and hopefully the next aortic stenosis patient gets a better examination.
Perry: Yeah, hopefully. Mine probably will get a better examination, too, after this discussion.
This episode was sponsored in part by Thinklabs, the creators of the Thinklabs One digital stethoscope.
is a cardiology fellow at the University of Washington Medical Center in Seattle.