LearnICU is a comprehensive library housing thousands of online educational materials, many available only to SCCM members. ​
Search the library or explore resources sorted by Knowledge Line (Category) or Content Type.

SCCM Pod-446 Comparison of Bivalirudin Versus Heparin

This webcast will be a comparative analysis of conventional heparin- versus bivalirudin-based systemic anticoagulation in adult and pediatric patients supported on extracorporeal membrane oxygenation (ECMO). Host Ashish K. Khanna, MD, FCCP, FCCM, is joined by Troy Seelhammer, MD, to discuss a retrospective study of adult and pediatric patients receiving ECMO from January 1, 2014, to October 1, 2019 (Seelhammer T, et al. Crit Care Med. 2021 Sept;49:1481-1492). Dr. Seelhammer is an anesthesiologist in the Department of Anesthesiology and Perioperative Medicine at the Mayo Clinic in Rochester, Minnesota, USA.

Published: 10/27/2021

Estimated Time: 27:35 min


Dr. Khanna: Hello and welcome to the Society of Critical Care Medicine’s (SCCM) iCritical Care Podcast. I’m your host, Ashish K. Khanna, MD, FCCP, FCCM. Today, I’ll be speaking with Troy Seelhammer, MD, on the article “Comparison of Bivalirudin Versus Heparin for Maintenance Systemic Anticoagulation During Adult and Pediatric Extracorporeal Membrane Oxygenation.” To access the full article, please visit ccmjournal.org. Dr. Seelhammer is an intensivist and anesthesiologist in the Department of Anesthesiology and Perioperative Medicine at the Mayo Clinic in Rochester, Minnesota, USA. Welcome Dr. Seelhammer. Do you have any disclosures to report to our audience today?

Dr. Seelhammer: No, no disclosures today.

Dr. Khanna: Thank you for your time. I know these are pressing times for the critical care community. We’re going to talk about a study that is focused on anticoagulation management and ECMO.  We all know how precious a commodity the ECMO circuit is, so clearly this is a very timely discussion. Would you like to start by summarizing what you did?

Dr. Seelhammer: Our study was a retrospective chart review of both adult and pediatric patients who required ECMO from January 1, 2014, until October 1, 2019. Our center at the Mayo Clinic is a large, high-volume tertiary referral center for both adult and pediatric ECMO. In total, we were able to include 424 individuals. About three-quarters of those were adults in both the bivalirudin and heparin groups. Of the smaller group, about 25% were in the pediatric groups. The purpose of the study was to compare the efficacy and safety of bivalirudin versus heparin-based anticoagulation during ECMO.

We found a variety of interesting things. Most importantly, I think—and this is the one that jumps out when people read this article the first time—is that there was a marked reduction in mortality in adults who received bivalirudin when compared to heparin. In fact, the reduction in mortality was substantial. The odds ratio was 0.39. I think that’s something we can spend some more time digging into and whether or not, after factoring everything in, this is something that is believable because it’s such a huge difference in survival. Some of the more nuanced findings were some trends, as you might expect, but not all of them reached cutoffs for statistical significance. But we did find that it seems there was no difference in the rates of bleeding and in most metrics, the rate of transfusion in both groups, which is a little bit different perhaps than some of the other studies that have been published in the past few years.

Dr. Khanna: This is clearly a very important question because all of us deal with heparin-based anticoagulation versus the alternative. Bivalirudin is certainly a very valuable alternative that all of us have utilized. What was your rationale when you started thinking about actually looking at these data? What made you think about this study?

Dr. Seelhammer: Well, like at most centers, unfractionated heparin was our workhorse, both where I trained at the University of Michigan and also at Mayo Clinic. It’s the go-to drug for systemic anticoagulation in mechanical circulatory support, primarily because it has the broadest breadth of experience. It’s available and it’s inexpensive and you can reverse it with a drug or protamine. However, it’s also very imperfect. It’s challenging to titrate and you have to worry about its cofactor antithrombin for its mechanism of action.

So you’re dealing not just with the drug effects but with fluctuating levels of this other essential cofactor. Lastly, heparin-induced thrombocytopenia, which, although not common in mechanical circulatory support, is found because these patients are exposed to heparin during their courses. If it were to develop, it markedly impacts the patient’s morbidity and mortality. So for many years there’s been a search for a better alternative. When I joined Mayo Clinic in 2014, we were managing a subset of individuals for whom we struggled to achieve that balance of adequate systemic anticoagulation and hemostasis, just from excess deposition on the circuit.

In a subgroup, we would employ a dual strategy where we would use argatroban, another direct thrombin inhibitor, in conjunction with heparin. The challenge there is that we now had two different drugs acting on similar pathways, and lab monitoring couldn’t discriminate the effects of each individual drug.

The addition of argatroban seemed to quiet down the laydown on our circuits. But because we had such a challenge with titration, it left us with a lot of questions, perhaps more questions than answers. So we looked to the literature for some guidance. And what we found were some case reports and some small case series and an article by Dr. Ranucci et al dating back to 2011 in 21 post-cardiotomy VA ECMO patients. They were able to find a reduction in the fluctuant nature of intensity of anticoagulation; that is, they were able to find that delicate balance and maintain it a bit better with bivalirudin when compared to heparin. And they bled less and they required less blood product transfusion.

So we decided to develop a protocol for the administration of bivalirudin in this population. Initially we targeted adults only. Once we developed a guideline based on what little evidence we had, we were able to employ it on selected patients where we felt like we were continuing to struggle with anticoagulation. And from there, things just slowly unfolded. We were able to mature our guidelines, and we can talk a bit more about that if that’s of interest.

Dr. Khanna: This is a fascinating story. A question that jumps out right away is: Within your study population, you chose to clump pediatrics and adults together. Was that done for a reason or do you truly believe that the interaction of heparin and bivalirudin in children and adults is the same when they’re on circuit or otherwise?

Dr. Seelhammer: That’s a very good question. We had had more struggles with anticoagulation in the pediatric population. The exposure of the artificial circuits compared to the circulating blood volume is much, much greater in pediatrics, particularly in the small neonates. Also, their anticoagulation systems or hematologic systems are not quite mature. We struggled historically with finding the right intensity of anticoagulation and finding a drug to help achieve that. When we had some success early with the argatroban-heparin combination and then, subsequent to that, some success with the bivalirudin protocol, we were very quick to roll out that protocol in an adjusted form to our pediatric population. Both groups had dedicated guidelines; the protocols were a little bit different but very very similar and nuanced just for the specifics that were relevant for the pediatric population. Having used this drug over the course of four years, when we started to develop this study to look back in time from start to finish when we first rolled it out and how things played out over that time interval, we felt that there was potential benefit in both groups.

Whether or not the overall efficacy or its impact is identical, I think that’s probably not possible for us to determine, but I do think that it was worth looking to see if indeed both groups could benefit from the intervention. We didn’t quite find that in our study. One of the rationales for looking at both pediatrics and adults is that it really hadn’t been done before, at least as it relates to bivalirudin, anticoagulation, and ECMO. So we wanted to get some evidence out in the literature to try and get the ball rolling on more sophisticated analysis in the future.

Dr. Khanna: The other thing that would be fairly obvious in an analysis like this is that you probably included patients over a certain period of time, and it looks like your results state that your inclusion rolls from January 1, 2014, to October 1, 2019. I’m sure that Mayo Clinic as an institution changed its anticoagulation strategies along the way. Do you feel that change over time also affected your results for this study?

Dr. Seelhammer: We had a lot of changes in our program prior to January 1, 2014, some of which were technological improvements. We had migrated to the CardioHelp system along with its associated sets and artificial oxygenator. We also had developed and rolled out an established protocol for the administration of unfractionated heparin that included a very tight time frame for initiation and titration of the drug, all guided by selected laboratory parameters. And we had rolled out a series or sequence of lab testing whereby, once hemostasis is gradually achieved, the frequency of anticoagulant monitoring is de-escalated.

That was all established prior to January 1, 2014. Additionally, we had changed from a historically pediatric intensivist-led ECMO program to a dual intensivist-led program whereby a pediatric intensivist would manage pediatric patients and an adult intensivist would manage adult patients. That all took place prior to January 1, 2014. So part of our rationale for that inclusion date specifically was to try to mitigate some of these confounders.

Additionally, during our analysis, we adjusted for year of ECMO. So we were trying to adjust for heparin patients who received heparin in a given year and trying to adjust for differences in the actual study year that was included in the trial, again, to try and adjust for things that we might not have measured, but there were no major programmatic or technological changes or other guideline changes in the study dates between 2014 and 2019. So yes, it is certainly possible, but nothing that we were able to determine to measure and include in a more sophisticated fashion.

Dr. Khanna: So what happens to the actual included patients? Your results say that you had 540 patients, of which you excluded 118. Then there were 89 pediatric patients, with some on heparin, some on bivalirudin, and 333 adults. Did you have to exclude a lot of patients because they were on mixed anticoagulation, which means that they started with heparin and at some stage they got heparin-induced thrombocytopenia, and then you switched to something else? Was that one of the major reasons for exclusion or otherwise?

Dr. Seelhammer: No, there was a small number that fell into that category, but we excluded a total of 118 patients. Of those, the majority were excluded because they didn’t receive systemic anticoagulation within the first 48 hours. So we didn’t feel that comparing that smaller cohort was appropriate because there’s typically a reason why we wouldn’t start anticoagulation during that time interval, so the population is potentially quite different. Then of course we had the typical patients who didn’t approve the research use of their medical record, and we had 25 patients in total who died within six hours of ECMO start. For obvious reasons, we didn’t include those either. The question though, about how to handle those patients who are shifted from one group to another, was one that we really struggled with, but we actually ended up having to pull our data three times and do three separate analyses to try and get things right. That’s one of the things that we really struggled with, both getting the accurate data poll or the abstraction piece right, and then how to analyze these in a fashion that was developed in advance of actual data collection.

What we tried to do is exclude the first initial bolus of heparin that they would have received, because that was often at the time of cannulation. And we would use the start of their systemic anticoagulation infusion as our initial categorization of their study group. A small number of patients—and it was a very small number—transitioned from the heparin group into bivalirudin. And those were suspected or confirmed heparin-induced thrombocytopenia. To my recollection, there weren’t any patients that were transitioned from bivalirudin into the heparin group.

Dr. Khanna: This is fascinating because you well realize that you’re sitting at a center where the use of bivalirudin as a primary anticoagulant is practiced. And I feel that that is certainly not the norm in a lot of other places, where it is still used as a rescue agent when things are not working well with heparin. Do you agree with that observation?

Dr. Seelhammer: Yes, that is very reflective of what has been published in the literature and what I’ve encountered in various centers and traveling around over the past few years.

Dr. Khanna: Do you feel that this really interesting data is now going to hopefully change clinical practice?

Dr. Seelhammer: I think it offers the possibility of practice improvement. I think it’s a bit too early to say with certainty that this should be used more broadly and at the majority of centers. But I do think, because of its mechanism of action, it’s easier to manage clinically. That by itself is a big selling point to centers that are considering the transition to bivalirudin. Frankly, by not having to worry about antithrombin, you don’t have to assay that particular compound and you don’t have to replace it. The titration tends to be easier too. Our data didn’t show a reduction in aPTT frequency, but we did markedly reduce the frequency of other assays, including no longer needing to use the anti-Xa assay. We didn’t need to use ACTs in any frequency. And we markedly diminished our dependence upon thromboelastography.

So although we didn’t change aPTT frequency, we probably dramatically reduced the overall lab frequency. And that’s something that we’d like to explore further in the future. At the bedside, when you’re managing this population, rather than trying to choose which assay you like and which lab results you don’t necessarily believe in that clinical context, now you’re left with a cleaner way of assessing the anticoagulate intensity and adjusting as necessary. In the future, this is likely to get even easier. At the Mayo Clinic, we are just rolling out our chromogenic factor IIa, which is an acid that is targeted and developed for the use of this drug class. My hope is that it offers an even better way of titrating the drug in the future. I think, because you can dose it so readily and adjust it so easily, it offers the clinician at the bedside an easier time with getting your head wrapped around this concept of finding that delicate balance between hemostasis and anticoagulation.

Dr. Khanna: All of this makes me think healthcare dollars. Did you plan to investigate an incremental cost-effectiveness analysis comparing not just the two agents, but it looks like it has significant cost saving in terms of not having to do as many anticoagulation tests in your bivalirudin group?

Dr. Seelhammer: The drug cost of bivalirudin is significantly more expensive on an equipotent basis versus unfractionated heparin. Depending upon your contractual rates, it’s at least 20 times the price for drug acquisition, which is even reflective of us buying much more of the compound since we administer it to so many more patients now, so the drug is much more expensive. So then the question is, how do you make up that difference? Can you make this cost effective for the total cost of administration? That would factor in things like lab monitoring, but also hemorrhage, needing to go back to the operating room when you’ve over- or under-anticoagulated your population, transfusion. All of these things can have an impact on the overall cost of administration. We had very much wanted to explore that in our study. We have some limitations at our center with the proprietary nature of cost analyses, which requires an entirely separate project and it’s something we fully intend to do in the future, but it did require a complete separate set of analyses at our center.

However, others have explored this. Dr. Hamzah et al at the Cleveland Clinic published a wonderful manuscript last spring, in early 2020. This was in the pediatric population and they specifically focused on a cost analysis. What they found is that the overall cost of administration of bivalirudin at their center for this population of pediatric ECMO patients was significantly lower than that of the heparin group. This is very reflective of what, going back to back in 2011, got us interested in this drug in the first place. Ranucci et al in 2011 also found cost-effectiveness in their population. So I don’t know if the verdict is out on whether or not this is indeed cheaper per se versus heparin, but I think it’s definitely a question worth asking. Those looking at further analysis using this compound in ECMO should definitely consider it as part of their analysis plan because I think it’s a question that we all really want answered.

Dr. Khanna: They say one good retrospective study sets the stage for several subanalyses and maybe a prospective randomized trial. Just jumping ahead, I’m sure there have been lots of suggestions on what to do next with these exciting data. What are some of the things that are high priority for you in terms of future directions with this data?

Dr. Seelhammer: One of our priorities going forward is to try to move this from a single-site to a multi-site approach. Although we haven’t yet broached the subject of a prospective trial, we have begun collaboration with several other large tertiary referral centers within the United States to try and pool our data. What we’d like to do is to, of course, increase our statistical power to try and better discriminate some of these harder-to-find-a-difference-in outcomes like circuit intervention rates, individual blood product transfusion rates, and further analyze the cost effectiveness. So we want to pool our data with other centers to try and get a better glimpse at those results.

Then of course we want to look at subgroups. We’re late in the process right now of analyzing our COVID ARDS ECMO patients to look at the efficiency or efficacy of bivalirudin in that cohort, because all of ours to date have received bivalirudin. This will be more of a descriptive series as opposed to a comparison series but, by collaborating with other centers, our hope is to try and compare efficacy as well at some point in the future.

One of the big things that our center is working on right now is reflective of one of the challenges that we had with this study, and that is data abstraction. We really struggled to get the data pool right. Hopefully we’re not entirely alone. And if we are, we’re going to try to do better in the future, but pulling all the nuanced data with ECMO is really difficult. Some things we weren’t able to pull at all, things like, what is our target anticoagulant level? We don’t have anywhere in our database what the goal aPTT is. It’s not documented. We have some daily notes that mentioned it, but the moment-to-moment goal is not reported anywhere. So we want to build a more robust dataset in the future where we can pull data more readily, more quickly, get the data to the larger research audience more quickly and try to make meaningful forward changes with our population.

Dr. Khanna: Yes, indeed, you hit the nail on the head. People get into doing smaller retrospective studies and you realize that you wish that the clinician, which could include us, had documented things in appropriate boxes where they could be extracted electronically, and that doesn’t happen all the time. I guess what you’re almost leaning toward is an establishment of a national ECMO registry, which is different from ELSO in that you’re focused on anticoagulants, for example, and you literally provide all your participant sites the exact data fields to enter and where, so you could get precise data as you build this registry.

Dr. Seelhammer: Yes, that’s part of our vision. We want to use the ELSO registry as a starting point, not because we want to necessarily replicate it, but we do want to try to figure out what’s been successful for the ELSO registry and what’s absent or missing from the registry in its current form. We’ll also look for other databases as well, like the STS database, and try and figure out how to do this the best way we can prospectively because whatever challenges that we can try to address now early on will make things much more fruitful later on. Our goal is to start with this as a Mayo Clinic enterprise-wide project, where we can roll in our other large ECMO sites in Jacksonville and Arizona, and then eventually broaden this perhaps to other collaborating centers in the future.

Dr. Khanna: Excellent. You touched a little bit on COVID-19. Clearly our audience would love to hear what your gut feeling about that is. You said you don’t have a control group on that, but say you did have a control group on that. What’s your gut feeling with COVID-19 and the behavior of anticoagulation? That’s one of the challenges as well, when we put these patients on ECMO.

Dr. Seelhammer: That’s the million-dollar question, to describe the way that we manage this aspect of this population’s care. We uniformly give a small dose of heparin as a bolus up front when going on ECMO. In all patients that we’ve managed today we institute a bivalirudin infusion early in their care, as soon as the hemostasis can be demonstrated and normalization of the aPTT toward normal range; it doesn’t have to fully normalize, just near normal. Then within the first 24 hours, we administer low-dose aspirin. And this is a time to pause and consider what’s different with the COVID population. We know that there’s a lot of laboratory derangements in this population. That’s been well publicized at this point. We also know that early reports of ECMO COVID ARDS have been fraught with thromboembolic complications, circuits that just continue to fail and require component replacements or full circuit replacements. That has complicated and challenged outcomes.

So, early on we decided we were going to try to leverage bivalirudin, and also leverage targeting primary hemostatic cascades with low-dose aspirin. Whether or not that’s been successful, I think we would really need to compare our cohort to a heparin-based cohort, but we have had quite favorable outcomes. Our survival rates are really quite good to date. So I’m optimistic that not just focusing on secondary hemostasis, but maybe broadening these to include primary hemostasis offers a clinical advantage.

Dr. Khanna: Excellent. Great results. There is a huge reduction in mortality in adults who received bivalirudin versus heparin. What all did you control for in terms of confounders?

Dr. Seelhammer: That’s a very good question: How do you control for all the different various aspects? One thing that’s abundantly clear with our dataset is that, by including all adults and all pediatric patients who met our criteria, we were including widely divergent patient populations, pathophysiologic states, even circuits and configurations. So comparing them head to head wasn’t really possible with our dataset. The population is just too diverse. That offers some advantages, but many, many disadvantages. We struggled with how to control for all this stuff.

So what we did try to adjust for, as I mentioned earlier, the year of ECMO that the patients were managed on. We also tried to include, of course, the age of the population in pediatrics, because comparing a neonate to an 18-year-old didn’t seem all that fruitful. We also included the cannulation approach—aortic versus peripheral—or central versus peripheral cannulation, and the ECMO type—VV versus VA. That’s really the extent of what we tried to adjust for as prespecified variables.

Dr. Khanna: This is fascinating stuff. And I’m absolutely certain that this is the beginning of a lot of new projects. In fact, I would say that this is probably great data to use to look for grant funding or to a large, much-needed prospective randomized trial that will clearly help answer this question, and that would probably be practice changing. Dr. Seelhammer, I want to thank you for your time today and sharing all of your great data with us.

Dr. Seelhammer: I appreciate the invitation to contribute. Thank you.

Dr. Khanna: This concludes another edition of the iCritical Care Podcast. For the iCritical Care Podcast, I’m Dr. Ashish Khanna.

Ashish K. Khanna, MD, FCCP, FCCM, is a staff intensivist and anesthesiologist, associate professor of anesthesiology, and a section head for research with the Department of Anesthesiology Section on Critical Care Medicine at Wake Forest University School of Medicine in Winston-Salem, North Carolina, USA. Dr. Khanna is heavily involved in SCCM specialty sections, committees, and task forces and has received the SCCM Presidential Citation multiple times. He has written more than 80 peer-reviewed articles and two dozen book chapters, as well as editorials, non-peer-reviewed articles, and online educational videos. He has been invited to talk about his work at prestigious national and international forums. His research interests include prediction of postoperative respiratory and cardiac events on the wards, outcomes of hypotension in critically ill patients, and the use of novel vasopressors in shock states in the ICU.

Join or renew your membership with SCCM, the only multiprofessional society dedicated exclusively to the advancement of critical care. Contact a customer service representative at +1 847 827-6888 or visit sccm.org/membership for more information. The iCritical Care podcast is the copyrighted material of SCCM and all rights are reserved. Statements of fact and opinion expressed in this podcast are those of authors and participants and do not imply an opinion or endorsement on the part of the SCCM, its officers, volunteers, or members or that of the podcast commercial supporter.

Some episodes of the iCritical Care Podcast include a transcript of the episode’s audio. Although the transcription is largely accurate, in some cases it is incomplete or inaccurate due to inaudible passages or transcription errors and should not be treated as an authoritative record.

Categories: Quality and Patient Safety, Pharmacology,
Content Type: Podcasts,