Aortic Pulsatility Index (API): Optimizing Cardiogenic Shock Management

Welcome to the blog post version of the Saving Lives Podcast. In this post, I’m going to be discussing the aortic pulsatility index. Chances are you’ve never heard of this before. The reality is that I honestly had never heard of this before either. I ran into it been going down my rabbit hole of optimizing my management of cardiogenic shock patients. Well, here we are.

First of all, for historical context, today is the 13th of May of 2023. The papers used as the citations for this post are all after 2021. There’s no mention of the Aortic Pulsatility Index before that. The other good news is that every paper used as a citation is listed below and is free to download. Therefore, you can double check my work as this is not medical advice. Before we get started, a definite hat tip to all the authors of these papers for their work. Because of them I could create this content. Let’s get started.

Current Management of Cardiogenic Shock

When we care of patients who have cardiogenic shock, there are metrics and hemodynamic parameters that we try to meet. We want to make sure that the patient is making urine, they’re normotensive. Hopefully they’re clearing their lactate. They’re warm and perfusing other extremities, etc, etc, etc.

In addition, there are also more advanced hemodynamic measurements that we look at these days. These include Cardiac Power Output (CPO) and Pulmonary Artery Pulsatility Index (PAPi). These are two different measurements that I have covered before on my YouTube videos, my podcast and on my website. To obtain those measurements, it is best to have a PA catheter (aka Swan Ganz) in place. Nonetheless, one of the things that we need to be academically honest CPO. CPO is not validated in patients with decompensated heart failure. To my knowledge, it is only validated in cardiogenic shock secondary to acute myocardial infarction.

So then the question is, what do we do to best optimize decompensated heart failure patients in cardiogenic shock? What’s the hemodynamic parameter or metric that is predictive for these folks? Is this where the aortic pulsatility index will hopefully come in and fill that void?

Patient Identification & Escalation of Cardiogenic Shock

Identification of patients with cardiogenic shock has always been quite challenging. I always have to tip my hat to the authors of the SCAI criteria for cardiogenic shock. This criteria helps us classify our patients on an A through E scale. I’ve discussed this in the past, but A are the patients who are at risk. B are the patients who are beginning to go into cardiogenic shock. C are those who are in classic cardiogenic shock. D are those patients who are deteriorating. E are the patients who are in extremis.

Keep that in mind as you take care of your patient with cardiogenic shock. Their prognosis is closely tied with their SCAI shock stage.

On top of that, when we try to take care of our patients, the first thing we usually do is start patients on either vasopressors or inotropes to buy some perfusion.

Obtaining Hemodynamic Measurements

Going back to the whole CPO and PAPI conversation that we were having just a few moments ago, one needs a PA catheter to help manage these patients and it’s something that I commonly harp on when I discuss this cardiogenic shock topic.

In the case of cardiac power output, we use a PA catheter to get us the cardiac output so we can plug it into the equation for CPO. Now, when we look at PAPI to help define what’s going on with the right side of the heart, you need a PA catheter because you need to calculate what the You need to obtain the numbers of the PA systolic minus PA diastolic, and you need to divide that by the right atrial pressure or the CVP.

For the Aortic Pulsatility Index there’s no difference here. We also need a PA catheter to go ahead and obtain these data to obtain this number to hopefully help us manage our patient who’s in cardiogenic shock.

Why Should We Use API Rather than CPO?

The first question I ask myself, and I’m gonna basically break this whole episode up into different questions that I’ve asked myself when going over these data is, why should we even use api? Why don’t we use a cardiac power output? And the reality is, as I mentioned before, cardiac power output is not validated in this patient population. And when I say this patient population, I mean those with decompensated heart failure.

One of the reasons why it is thought that we should use the aortic pulsatility index, is that it “simultaneously represents cardiac function as well as filling pressures”. This was stated by Belkin and his team, which again, these are the pioneers reporting the first data regarding API.

In a different paper, Belkin and his team looked at patients who were enrolled in the ESCAPE trial. To refresh people’s memories as to what the ESCAPE trial was, is that was patients who were in decompensated heart failure and they received a PA catheter to help them be managed or standard care. One of the limitations key limitations of the ESCAPE trial was that they did not enroll patients who were in Cardiogenic Shock. Therefore, when Aortic Pulsatility Index data is extrapolated from the ESCAPE trial, just know that it can’t be used for Cardiogenic Shock.

Outside of the ESCAPE trial, there was a different trial where they also used retrospective data from the cath lab who were having milrinone stress tests performed. Measurements were obtained including the different components that are needed to actually calculate the aortic pulsatility index, which I’m going to get into the calculation in just a second. Well, actually, let’s get into it now.

How to calculate the aortic pulsatility index (API).

Aortic pulsatility index is equal to systolic blood pressure minus diastolic blood pressure over the pulmonary capillary wedge pressure. This is a pretty simple equation, but again, you need the PA catheter in place to actually complete the equation.

Otherwise, how are you going to get that wedge pressure? I am aware that you could go ahead and get an LVEDP when the patient’s having a cardiac cath performed. Although it’s not practical to be doing this at the bedside to measure what’s going on with the left side. Hence the pulmonary capillary wedge pressure is the way to go.

Oscillometric BP Cuffs vs. Arterial Lines to obtain the SBP and DBP in API

Now, a basic qualm, which I hope that we’re able to sort out in the future with other studies, is… How is it that we are obtaining the systolic blood pressure and the diastolic blood pressure to Plug into these this equation? The reason why I ask that is a topic I have discussed in the past. If we rely on oscillometric blood pressure cuffs, then we have to accept that the most accurate number of the three is the MAP. I’ve discussed this before and here’s a link to my POST on this topic. Instead, when you use an arterial line, you are measuring the direct pressure. But then, you know, when you have an arterial line, are you validating it based on a radial line, a brachial line, an axillary line, a femoral line? Those are all questions that. We need to think about when it comes to plugging in the numbers into the aortic pulsatility equation.

What Cutoff Should Be Used for Aortic Pulsatility Index?

So after these teams did these statistical jumping jacks, well, the question is, what cutoff should be used? What API values should tell us that the patient’s not going to do well? Based on these Belkin data, it was noted that 1.45 is the number to shoot for.

If the API is greater than 1.45, the patient’s chances of doing well, which were defined as freedom from advanced therapies or death was 79%. If the API was less than 1.45, that number was just 48% at 30 days. Not good at all.

Now those were patients who are in cardiogenic shock, but what about the patients who are not in cardiogenic shock? And this was looked at in a different study. Again, all these, all these data are linked in the show notes below. And I definitely recommend that you read these for yourself and not trust me because again, this is not medical advice.

Here they found that the cutoff to reach this composite endpoint of death or need of heart transplantation or LVAD at six months was greater than 2. 9, okay, greater than or equal to 2. 9. So you might say, okay, that number is a lot higher than, you know, the one that, the one that I said before of 1. 45 and yes, this is true, but it’s not even the number that’s kind of.

Making me question this a little bit more. It’s also the area under the receiver operating curve. Okay, and for those of you who are not statistical, uh, nerds like, like I am, basically what the area under the curve is it plots the sensitivity and specificity of a test and then, you know, tells us how good it is.

And for the sake of context, An area under the curve of 0. 9 and higher is considered to be like the gold standard types of tests, right? And there’s, there’s an article that kind of explains this, but if your area under the curve is 0. 9 or greater, the test is amazing, fantastic gold standards type stuff.

If the area under the receiver operating curve is between 0. 8 and 0. 89, for example, then this is still an excellent test, right? But if it’s between 79, then it’s a, it’s a pretty good test. It’s not, not something that’s excellent. It’s not something that’s going to be the gold standard. But anything less than 0.

69, 0. 69 and below, it’s kind of like a coin flip. It’s not too sensitive, nor too specific. It’s just not the type of test that you want to base your practice on. And here it turns out that the area under the receiver operating curve for predicting that endpoint was just 0. 71. which is not something to write home about, okay?

It’s not, and this is the reason why, um, I don’t think that the API is ready for primetime. When they looked at CPO, for example, in the same patient population, you know, those with decompensated heart failure, the CPO, uh, area under the receiver operating curve was 0. 69. So, um, you know, that’s, that’s almost as predictive.

Excuse me, the area under the curve was 0. 58. 0.69 was the threshold that they used. So. In CPO, I apologize. In CPO, if the CPO is less than 0. 69, the sensitivity and specificity of a bad prognosis was just 0.58 for the area under the receiver operating curve. Again, that’s almost as predictive as a coin flip.

So to sum up those points, I guess we could say that an API that’s greater than or equal to 1.45 in patients with cardiogenic shock means that they’re going to do okay. If the API is greater than or equal to 2.9 in patients with decompensated heart failure, they’re going to do okay.

More recently in 2022 Belkin et al used a “validated computer simulation model, model to simulate pressure volume loops”. They ran model on patients with acute heart failure as well as acute myocardial infarction. They looked at difference SCAI shock stages. All in all, the APIs for these different groups, were quite different depending on the shock stages.

The truth is that I don’t really know what to do with those data. I definitely tip my hat to the authors, but I would really like to see that work done on actual individuals who have, you know, they’re not just models, right? They have complexities and they have other variables in their body that just can’t be run on a computer simulation.

There have been attempts to improve the validation of API. Siddiqui et al. looked at combining API with PAPI, the Pulmonary Artery Pulsatility Index. This was in an attempt to strengthen the prognostication of this tool. It seems like there might be some sort of benefit to doing this. Then again, this is all retrospective data that has not prospectively been studied/validated.

Has the Aortic Pulsatility Index been evaluated on Vasopressors?

The short answer is no, it really hasn’t. Definitely not on vasopressors. Inotropes, yes. I had mentioned earlier that the first retrospective Belkin studies were giving patients Milrinone. Those of us familiar with milrinone know what it does to the SVR. That’s just something that we don’t know right now.

Has the Aortic Pulsatility Index been evaluated on Acute Myocardial Infarction?

The short answer here is also no. When you look at the Belkin studies with the Siddiqi studies, they only included patients with acute decompensated heart failure. No patients with acute MI have been evaluated for this. At least we have data that we should be using CPO for that patient population.

What other uses do we have for API?

Yes, there’s some data from Sanchez et al. looking at the utilization of API to possibly diagnose severe aortic regurgitation. That’s not something I’m going to go into in this post because we’ll be here forever.

Can we use the Aortic Pulsatility Index for Management Decisions Today?

My short answer is no. All the data we have is retrospective. Although it was all published in the last 2-3 years, the data comes from registries. Those registries go back to 2013. Our management of patients with cardiogenic shock has changed significantly over the course of the last decade. The advent of mechanical circulatory support devices, as well as protocols that have helped us earlier identify these patients. We have improved our game when it comes to patient management.

We can’t use these API, in my opinion, because it has not been validated prospectively nor robustly. In addition, we’re seeing an area under the receiver operating curve of just 0.7x. That’s disappointing. It’s better than a coin flip but not something I’d necessarily hang my hat on. Hopefully we will be able to use API to be able to prognosticate and help us manage our patients. For example, should we escalate the patient? Could we wean down the devices?

Do we need to do that? It’s potential right now. It’s just limited. Hopefully data will come out soon that will help us better utilize this marker to this hemodynamic variable to prognosticate our patients and ultimately take better care of them.

So where do we go from here? And other people looking at this, I know I can’t use my particular institution because we’re not wedging our patients who are in cardiogenic shock. At least I’m not routinely in my clinical practice, but I’m sure that there are institutions out there that have a breadth, a huge amount of.

Data that they could look into in their EMR of patients with cardiogenic shock who have a PA catheter who have received it You know, I’ll be, excuse me, who have received pulmonary capillary wedge pressure Readings, you know routinely to calculate what the API is and then kind of sort out what the outcome was of that patient Hopefully that type of data could be published soon so that we can have a better idea and then hopefully start a prospective study or at least You know, do a sub analysis of a prospective study of a patient group of cardiogenic shock patients from acute decompensated heart failure.

And then, you know, hopefully give us some numbers that we can actually work with and, you know, help us manage the patients because we need all the help we could get. The other caveat is, of course, in order to obtain these data, you do need a PA catheter. Um, I know that there are other ways of kind of estimating what’s going on with the left ventricle that could possibly be used as a denominator.

But you know, that’s the best we have now and definitely a hat tip to all the authors for all their hard work in trying To solve this problem. Hope you guys enjoyed this podcast this youtube video However, you get this content because it’s been a lot of fun to create. Thank you for your support. Have a great day.


Belkin MN, Alenghat FJ, Besser SA, Nguyen AB, Chung BB, Smith BA, Kalantari S, Sarswat N, Blair JEA, Kim GH, Pinney SP, Grinstein J. Aortic pulsatility index predicts clinical outcomes in heart failure: a sub-analysis of the ESCAPE trial. ESC Heart Fail. 2021 Apr;8(2):1522-1530. doi: 10.1002/ehf2.13246. Epub 2021 Feb 17. PMID: 33595923; PMCID: PMC8006667.
Link to Article

Belkin MN, Kalantari S, Kanelidis AJ, Miller T, Smith BA, Besser SA, Tehrani D, Chung BB, Nguyen A, Sarswat N, Blair JEA, Burkhoff D, Sayer G, Pinney SP, Uriel N, Kim G, Grinstein J. Aortic Pulsatility Index: A Novel Hemodynamic Variable for Evaluation of Decompensated Heart Failure. J Card Fail. 2021 Oct;27(10):1045-1052. doi: 10.1016/j.cardfail.2021.05.010. Epub 2021 May 25. PMID: 34048919; PMCID: PMC9073373.
Link to Article

Combined Utility of Aortic Pulsatility Index and Pulmonary Artery Pulsatility Index for Risk Stratification in Advanced Heart Failure Patients. U.A. Siddiqi, M. Belkin, S. Kalantari, A. Kanelidis, T. Miller, N. Sarswat, A. Nguyen, B. Chung, G. Kim, B. Smith, V. Jeevanandam, S. Pinney, J. Grinstein. The Journal of Heart and Lung Transplantation/ Volume 40, Issue 4, Supplement, April 2021, Pages S247-S248.
Link to Article

Belkin MN, Shah J, Neyestanak ME, Burkhoff D, Grinstein J. Should We Be Using Aortic Pulsatility Index Over Cardiac Power Output in Heart Failure Cardiogenic Shock? Circ Heart Fail. 2022 Jul;15(7):e009601. doi: 10.1161/CIRCHEARTFAILURE.122.009601. Epub 2022 Jun 6. PMID: 35658463; PMCID: PMC10038120.
Link to Article

Deis T, Rossing K, Gustafsson F. Aortic Pulsatility Index: A New Haemodynamic Measure with Prognostic Value in Advanced Heart Failure. Card Fail Rev. 2022 May 16;8:e18. doi: 10.15420/cfr.2022.09. PMID: 35620383; PMCID: PMC9127634.
Link to Article

Sanchez F, Goudelin M, Evrard B, Vignon P. The Aortic Pulsatility Index: A New Sign of Severe Acute Aortic Regurgitation. J Am Soc Echocardiogr. 2022 Sep;35(9):1006-1007. doi: 10.1016/j.echo.2022.04.008. Epub 2022 Apr 27. PMID: 35487473.
Link to Article

How to Support My Work

Discover “The Vasopressor & Inotrope Handbook: A Practical Guide for Healthcare Professionals,” a must-read for those in the field! You have several options to get your hands on a copy while supporting my endeavors. If you’re in the US, you can order directly from me with a special touch: a signed copy. Please note that I handle these shipments personally, so I appreciate your patience.

I recommend purchasing through THIS AMAZON LINK for quicker delivery or international orders. When you use this link, I earn an additional commission at no extra cost to you, which is a great way to support my work.

Stay tuned for the audiobook version, which will soon be available on Audible. It’s an exciting addition that I’m eager to share with you!

I have started using Descript to edit my videos rather than just Final Cut Pro. I have found it to be extremely helpful at saving me time. CLICK HERE to give it a go.


Although great care has been taken to ensure that the information in this post is accurate, eddyjoe, LLC shall not be held responsible or in any way liable for the continued accuracy of the information, or for any errors, omissions or inaccuracies, or for any consequences arising therefrom.