“The Pediatric Critical Care Transfusion and Anemia Expertise Initiative” S. Bateman, S. Valentine

“The Pediatric Critical Care Transfusion and Anemia Expertise Initiative” S. Bateman, S. Valentine


Welcome to World Shared Practice Forum. I’m Dr. Jeff Burns, Chief of Critical Care
at Boston Children’s Hospital and Professor at Harvard Medical School. We’re very pleased to have with us today Dr.
Scot Bateman. Dr. Bateman is Chief of the Division of Critical
Care Medicine and Professor at the University of Massachusetts Medical School. Along with us, we also have today his colleague,
Dr. Stacy Valentine. Dr. Valentine is Associate Professor of Pediatrics
and Associate in Critical Care Medicine at the University of Massachusetts Medical School. Scot, Stacy, welcome. Thanks. It’s good to be here. When I use hemoglobin concentrations, which
I’ll be using throughout the talk, I’m going to be using the conversion that we use in
North America, which is grams per deciliter. So if I say 8, that’s 8 grams per deciliter. So the SI conversion for that is grams per
liter. Dr. Bateman and Dr. Valentine, you both are
the authors of many of the recent guidelines in the last year or two on the appropriate
management of red cell transfusions in a critically ill child. And I wonder if we could begin by asking,
could you summarize the salient literature to date on what people should know that led
to the work that we will discuss in a bit. But what should people know about red cell
transfusions in a critically ill child? Sure. I think one of the things that we’re trying
to stress through this process you’ll hear about is that there really isn’t much literature
to support. In fact, there’s a paucity of literature specifically
on red blood cell transfusions in pediatrics, in pediatric critical care specifically. So what we did was try to take what we knew
and apply it and actually see how we could expand it. And that’s what we’ll try to explain today. So I can give a little bit of the background
just to give you a sense of some of the pertinent data that we have that really formulated where
we started from, and then we can go into what process we put into it. So I can start with that. I will say we have no conflicts of interest
to report, which I think is important for us in this process. We did get funding from the NICHD and NHLBI
through a conference grant to help bring these folks together. We also got support from the Washington University
Children’s Discovery Grant, the Canadian Institute of Health Research, and the Society for the
Advancement of Blood Management. We had support from the Pediatric Acute Lung
Injury and Sepsis Investigators (the PALISI), and the subgroup of that called BloodNet,
which was a very active part of our group. The Society for Critical Care Medicine backed
this, and also all of our T.A.X.I. experts, which T.A.X.I. was the term that we ended
up using which stands for Transfusion and Anemia Expertise Initiative. So when we mention T.A.X.I., that’s what we’re
talking about. We started this process looking back at the
most sentinel article in Transfusion and Critical Care Medicine, which came out in 1999 by Paul
Hébert, which is in the New England Journal of Medicine, where it was the first study
using transfusion thresholds. The standard threshold had always been 9 or
10 as far as the hemoglobin for giving a transfusion. And he published an article in approximately
1,000 adults looking at using 7 and 1/2 versus 9 and 1/2 as a transfusion threshold, which
really was revolutionary to us and our thinking. And really, I remember that article coming
out because it was the first article that really questioned our standards of care. And really, we hadn’t had anything to say
that maybe we shouldn’t be giving transfusions as much as we have. And I think at the same time, there was data
coming out that there were complications of transfusion that we kind of took for granted. There was always the infectious complication
with transfusion, but we also had inflammatory effects of transfusion, whether the transfused
blood cells were actually working. With the longer storage of red blood cells,
it may be not as effective. It really started the conversation going that
maybe we should be really looking at our transfusion practices. So that study really started the process. There was some data to suggest what are we
doing in pediatric critical care, because that was– all of that was in adults, and
it was really a watershed time for adult critical care. But we really we were a little behind in pediatric
critical care. So there was a survey done in 2002, which
is a couple of years later, looking at using different clinical states; bronchiolitis,
septic shock, trauma. And in a survey, what do pediatric critical
care practitioners transfuse by? What do they use as a threshold? And the data suggested that our transfusion
threshold in our pediatric patients was much higher, like 7, 8, 9, 10, more like an average
of about 8 to 9. And this was obviously very different from
what was being used in the Paul Hébert trials, called the TRICC trial, and so that was much
higher. And that was really the standard, and I was
practicing at the time, and that was what we did. We just transfused at 9 or 10. That was just the standard of care. And that survey really supported that practice. And then I was lucky enough to get involved
with a large, multi-center prospective study looking at our transfusion practices and the
incidences of anemia in pediatric critical care, about 30 centers across the US and Canada. And we found that actually, there is a big
burden on our pediatric patients. So it was very relevant to us how we transfused. So we found in patients who were in the ICU
longer than 48 hours, 49% of the patients who were in that period of time got transfused. And we also found that anemia is a big burden
as well. 33% of those patients were anemic on admission,
and another 41% became anemic in the hospital. One of the things that got Stacey and I together
is that she, as a resident, came to me. And she was interested in the development
of anemia and the influence of phlebotomy on the development of anemia in the pediatric
critical care area. And we were able to do subanalysis of this
data to show that phlebotomy was a significant predictor of anemia and transfusion, that
our practice of what we did to the patients in that PICU actually mattered. We also found that the pre-transfusion hemoglobin
threshold of the patients in the study was 9.7, so it was clearly much higher. And this was in about 2005, 2006, so up to
seven years after the TRICC trial had come out. And when we looked at what it was that triggered
people to give transfusion, what was the reason, low hemoglobin was, by far and away, the most
common reason, which would make sense. But then so it really argued that how we defined
thresholds for transfusion was very critical in how we were going to manage how to guide
people in doing transfusions. So, also at the same time was a study that
is really the seminal study in pediatric critical care called the TRIPICU study done by Jacques
Lacroix from Montreal. It’s the Transfusion Strategies for Patients
in the Pediatric Intensive Care Unit, called the TRIPICU, where he looked at a very similar
methodology to the TRICC trial done by Hébert looking at using a transfusion threshold of
9.5 versus 7. So if you had patients within their restrictive
group, which was the 7, then you did not transfuse unless you got below 7, and with a target
hemoglobin of about 8.5 to 9.5. And if you were in the liberal group, you
got transfusion if you were less than 9.5 for hemoglobin with the goal target post-hemoglobin
of 11 to 12. The caveat of the study was it was done in
a wide range of pediatric intensive care units across, actually, North America and Europe. But the patients had to be hemodynamically
stable, so it was a very specific subset of patients in the pediatric ICU. So hemodynamic stability was determined as
not needing a change in– you could be hypotensive, but you couldn’t have changed your blood pressure
medications, or your had no pressers at all. So there was an important piece to put in
so that– because there was still a sense at the time that transfusion in a hemodynamically
unstable patient was probably not willing– ready for prime time as far as a randomized
control trial. And in fact, we were approached at Boston
Children’s to be a part of that original study, the TRIPICU study. And we weren’t that comfortable with going
into that study because at the time, it felt like it potentially was putting our patients
at risk. That’s how far this field has come since then,
because at that time, we were still very wedded to the 9 or 10 as a hemoglobin threshold. So it’s been interesting to see how that has
evolved. So that study, the TRIPICU study, really did
demonstrate that there was no difference between the restrictive strategy or the liberal strategy
of transfusions. So there was– it was a non-inferiority trial,
meaning that there was no worse outcomes. There was no new or progressive multi-system
organ dysfunction syndrome, no increased risk of mods with severity of illness for the restrictive
group. And importantly, there was 44% fewer transfusions
in the restrictive group than there were. The study that I was involved with, the observational
study, did show that there actually was some– not a mortality difference, but there was
some morbidity difference in the transfused patients versus non-transfused patients, indicating–
and that is consistent with adult data suggesting that there is some other potential morbidity
related to transfusions in critical illness that we need to address and need to look at
much more systematically. And so it really has solidified our take that
avoiding unnecessary transfusions is really the key to why we did this in the first place. Because I think that being smart about how
we transfuse pediatric patients, I think, is something– is our goal. Stacy actually followed up her phlebotomy
work to look at how our transfusion practices have changed since prior to the TRIPICU study,
then following a couple years after. And there has been a gradual reduction in
the hemoglobin threshold that has been used in pediatric critical care, but it still is
higher than what is recommended in adults. And still, we felt like there was still room
for improvement in our sense of avoiding unnecessary transfusions, which led us to the point of
saying, OK, where do we go from here? And Stacey and I discussed this. We work together, and we had been involved
with transfusion research. Up to that point, we’d provided some of the
background data. And we felt like there was not enough data,
but there was still room for improvement in how we transfuse. And Stacey had actually just participated
in a Pediatric Acute Respiratory Distress Syndrome Consensus Recommendations, which
is called PALICC, where they had used a group consensus process to try to create guidelines
and criteria for pediatric acute lung injury. And we were talking in the hallway and said,
we should do this. In fact, she said we should do this for transfusion. I said, you’re absolutely right. We should. And then we made a pact that we would do it
together since that time, and here we are. We’re still doing it. Here we are. And I’m going to turn it over to her because
she really helped put together our process really well, and it’s something we’re particularly
proud of. Thank you. Scot, that’s a wonderful overview of the literature
of, really, the last 20 years. As you noted, I remember being in the conference–
in the conference room, and I had concerns about whether we should join Jacques Lacroix’s
study. And I didn’t feel comfortable withholding
packed red blood cells. Now, I’m asking myself– and now I’m asking
you– you’re the expert– what was the central hypothesis that drove the trial? What was it that they were questioning? We were coming out of the age of normalizing
values. We learned in mechanical ventilation that
high tidal volumes were not good. Was it similarly that people were questioning,
are we normalizing values when we shouldn’t, or was there emerging evidence that it was
impacting somehow the inflammatory response, leading to worse outcomes? What drove that research? Yeah, I think that’s a great question, because
the data really—what’s amazing to see, looking back, how little data there was to
show the benefits of transfusion in the critical care realm. And I think there was an assumption and an
extrapolation historically from oncology or bone marrow failure transfusions. And so there was a sense that critically ill
patients were anemic, so they should get transfused just like an oncology patient who may not
be able to make red blood cells. But what’s, I think, really the impetus to
the original TRICC trial was that critical illness anemia is different. And I think that the inflammatory state of
the body is different in critical illness and that red blood cells that have been stored
and donated and given to a foreign person– I think there was emerging data that it actually
may not be that helpful. And we may actually potentially be harming
them. And we should really be thinking about this,
because we really didn’t have the data to suggest that it’s really making a huge impact. It always– we always felt like it probably
did. And we’re improving oxygen-carrying capacity,
we’re improving their volume, but we were also introducing a foreign inflammatory marker
into the body during critical illness. And I think– if anything, the impetus, I
think, was really that potentially, the benefits from red blood cells may actually not be as
great as we anticipate or have always assumed. And we really need to say, does transfusion
actually improve outcome? And if we don’t transfuse as much, is it the
same, better? And I think that that period between 1999
and when we started this process, we found that actually, the data was suggesting that
lower was actually fine, maybe even better. And so we should really capitalize that and
try to help our colleagues in pediatric critical care know what is the best transfusion practices,
because there really wasn’t enough guidelines out there for any of us. And so we felt like we needed to do something
to bring the experts together and make some thoughtful and evidence-based recommendations
if we could. And so now the story turns to you, Dr. Valentine. And before you pick it up, could you go back
a little bit and tell us– Dr. Bateman alluded to the fact that you approached him as a young
resident, but what was it that drove your interest in this research? Yeah, I was a second-year resident at Boston
Children’s. Dr. Bateman was my attending. And I had taken care of a number of bone marrow
transplant patients who were very anemic and were receiving blood cell transfusions, and
I couldn’t help but notice that with phlebotomy, that we were taking more and more blood. And could we be contributing to this very
anemia that this patient had? And I just began to develop an interest in
anemia, phlebotomy, identifying factors that we could potentially help and guide transfusion
practice. And that really just started the ball rolling. And as I dove deeper and deeper into transfusion
practices, I became more and more interested in why we’re doing what we’re doing. And could we improve what we’re doing? And how could we, through research and developing
more evidence, bring people together to be able to provide that research to improve practice
as well as to provide guidelines to be able to improve practice? And it started then and has continued many
years later, and through the mentorship of Dr. Scot Bateman and all of my colleagues. But it’s really that passion for being able
to improve transfusion practices in critically ill children. That’s a great story. So now the story picks up. Yes, through this process, as Dr. Bateman
had mentioned, I had been on PALICC and was very interested in bringing this to the transfusion
world. So what I’ll do now is outline the approach
of creating consensus of clinical blood cell transfusion decision-making for critically
ill children. The Pediatric Critical Care Transfusion Anemia
Expertise Initiative; the aims are to create evidence-based– and when evidence is lacking,
expert-based– consensus recommendations on blood management strategies for clinicians
caring for critically ill children. Aim to maintain a physiologically relevant
hemoglobin as well as optimize hemostasis and minimize blood loss. Through this process, we took a staged approach. So red cells is only the first stage of T.A.X.I. And we decided to focus on red cells first. It was a three-part consensus series. It involved international multidisciplinary
experts on red cell transfusion and represented the pediatric fields of critical care, cardiology,
transfusion medicine, hematology, oncology, surgery, and anesthesia. We engaged experts on guideline development
and implementation science from the beginning, and it was modeled after the PALICC methodology. Next, I’ll go through the process of T.A.X.I. It began in October of 2014, where we prepared
for and created an organizing committee. We, at that point, defined the methodology. We selected our topics and selected our experts. The organizing committee was comprised of
Dr. Bateman and myself, and the executive committee was comprised of the executive committee
from BloodNet and provided oversight during the entire process. Our 49 experts were from eight different countries
and 29 academic institutions listed here– and we have to thank our T.A.X.I. experts
for all of their contributions to this– and again, represented those fields of pediatric
critical care, cardiology, hematology, oncology, anesthesia, surgery. We thought it was very, very important to
have a multidisciplinary approach when thinking about the critically ill child. And then also had implementation experts as
well as the evidence-based medicine experts. During that first meeting, all of the experts
convened. And at that point, we discussed the methodology. We finalized it, discussed the subtopics,
and developed our PICO questions. It was very important for our evidence-based
medicine experts to be able to provide their inputs, and we followed the Institute of Medicine
standards in creating guidelines. We also had our implementation experts at
that very first meeting to talk about when you’re creating recommendations or guidelines,
how do we think about implementability from the very beginning before we even develop
the recommendation, which is very important. So there are actually criteria for implementation
called the GLIA tool. And it’s a guideline for implementability. And in this, our implementation scientists
could actually work with the experts as they’re developing their recommendation– not to change
the content, but to think about the wording very purposely to make sure that these recommendations
could then be implemented at the bedside, which is the entire purpose of the process. During that first meeting, we finalized the
subtopics. We had nine subtopics in total. We first wanted to look at an overview, so
the indication for red cell transfusion based on hemoglobin and physiologic thresholds. Where were we at this state? And then population-based, so looking at transfusion
thresholds in the acute brain injury patient in congenital cardiac disease, sickle cell,
oncologic disease, respiratory failure, shock, life-threatening and non-life-threatening
bleeding, extracorporeal support, dialysis, ventricular assist devices, and alternate
processing of blood products. We developed our PICO questions at that meeting. The bulk of the work happened between the
first and second meetings where we did a systematic review, a very comprehensive literature search
using our evidence-based medicine experts in our library and sciences. And we performed those searches. And with that, each of the experts from those
subgroups reviewed the abstracts and included manuscripts. They were all reviewed by two, and then any
conflict resolution was performed by a third reviewer. We also felt it very important to use the
GRADE methodology to be able to grade the evidence. So that happened during that first and second
meeting as well. And then with that entire process, we developed
what we call our short text recommendations. So each subgroup, each expert developed those
recommendations, and again, with the guidance from the evidence-based medicine experts as
well as the implementation scientists. At that point, we reconvened at our second
expert meeting. And during this meeting, we presented each
of the short text recommendations. It was very, very important. Some of the recommendations were discussed
for two, three hours because it was very, very important to make sure that we were deliberate
in what we were intending to recommend. We wanted to ensure the draft recommendations
were clearly worded, were unambiguous and easily understood by each of the experts. The implementation scientists were again helping
us and looked at each of the recommendations and graded them based on the GLIA tool. And the recommendations were revised until
we were able to reach agreement during that meeting. Between the second and third meetings is when
we started to vote. It was anonymous. And we used the RAND UCLA appropriateness
method. And during this, there are three rounds of
online scoring. All the experts were encouraged to vote. And you would vote between disagreement of
one, or agreement of nine. And if you had a reason for disagreement,
if you were equivocal or disagreed, we encouraged to write that reason for disagreement. Any recommendations that had any disagreement
were sent back to the subgroups for potential rewording. The a priori agreement that we decided was
80%. Once that had happened, those two rounds of
voting had occurred, we met for our third meeting, and then our short text recommendations
were discussed. At that point, all the recommendations had
met agreement. At that third meeting, we then discussed the
recommendations. We discussed those fine-tuning of wording. Even though they had already met agreement,
if we made even the slightest change in the recommendation, we sent that back for voting
again, because we wanted to ensure that that change didn’t alter the meaning of the recommendation. With that, the third round of voting had occurred,
and all recommendations had reached agreement. During this third meeting, we also talked
about implementation strategies. How are we going to take these recommendations
and be able to bring them to the bedside? What tools will we be able to use? Would we be able to create a decision tree,
computerized decision support, other tools? And that’s where our implementation scientists,
again, brought forth identifying barriers. How could we help with physician behavior? How could we actually change potential transfusion
strategies, and what were the barriers? And then the last part of that meaning was
discussing the knowledge gaps, because during T.A.X.I., we also, as importantly as present
the evidence, also present the areas that lack evidence. And how can we put forth research priorities
so that we can then focus our efforts in terms of future research in transfusion? After that third meeting, we created the transfusion
decision tree and worked very closely with Pediatric Critical Care Medicine, which we
have to thank their tremendous support during this entire process. And we published our recommendations in Pediatric
Critical Care Medicine in 2018. And we’re most proud of that decision tree
that really can be put up at the bedside to really help the decision making, walking a
patient through that entire process, which we will highlight now. That’s fascinating. So the inclusion criteria is that they had
to have demonstrated that they did research in the area– that is, specifically on critically
ill children. But can I press you again? Tell me again where you got the funding. You said you got it from the NIH, and you
wrote what’s called an R-13. Over 90% of our experts attended every conference,
which is a tremendous amount of effort. And it’s really critical to have them in person. You can do it on phone, but it’s very different
when you’re in person discussing, that back-and-forth that you have, watching their face as they’re
watching: there’s so much more than just words when you’re trying to create consensus. And it’s really, really important to have
everyone in the room to make sure that everyone is in preliminary agreement for the recommendations,
which is why we felt so passionate about having them in person. I learned that from PALICC, from Philippe
Jouvet. We certainly learned a lot through the way,
and really, just with that ultimate goal to be able to provide that guidance at the bedside,
and so really taking that 100,000-foot view and saying the reason why we are here is to
be able to help that critically ill child. How do we make these recommendations such
that they improve practice but also can be implemented? Because we need to make sure that the practitioner
can actually take this and implement it at the bedside. If it’s not able to be implemented, then we’re
not actually making change. Everyone was there and passionate about improving
practice. And to have this group of people in the room
to say we can make a huge difference for critically ill children was very inspiring. And also, just– we realized also that we’re
here, and we need to make this happen. We need to have the best product possible. Yeah. I would echo that. I would say that there was clear vision among
the people who were able to recruit into this process. And I think that the buy-in we got from the
experts who really– in fact, Jacques Lacroix, who’s one of– did the TRIPICU study, he started
off saying, you know, I’ve always wanted to do this. This is so needed. And it really just set the tone for someone
of that level to say this so needs to happen, and we’re going to make it happen. And the process, like Stacey said, was great. We were amazed that everyone came. The first meeting, it was– we had over 40–
I think it was 45 people at most of our meetings. And I think the process was there was a lot
of recommendations. We didn’t make it to our final. There’s a lot of things we really wanted to
be able to provide guidance on that we couldn’t. I think the ones we couldn’t put in there
really fueled the knowledge gap aspect of our care. But what we did finally come together with
and make recommendations on, we all felt strongly about. And so I think it’s– that process left us
very confident that what we did put out there was something that we all felt very strongly
about. What a helpful roadmap for colleagues across
the world on how to develop a process to be thoughtful about expert consensus, as you
just said, Scot, to fill knowledge gaps. That was wonderful. Well, can we turn now back to the specific
topic? Yeah, sure. And so where does the story go from here? Yeah, so this is the fun part. We can actually talk about the recommendations
themselves and the process. Last year, in September of 2018, we had a
full supplement that outlined all the different subgroups and the major recommendations that
came out of them in the background. And at the same time, partnered with that
was a special article that outlined all of our recommendations and that we’re going to
go through piece by piece here for our listeners. We ended up with a total clinical recommendations
of 56 recommendations that ended up being included, 45 specific research recommendations
as well. We had a wide range of types of quality of
the recommendations, so we graded the recommendations. So as I mentioned before, there was really
not a lot of data for many of our recommendations. But we did have– 36 of the recommendations
were based on the expert consensus from our experts. Of those recommendations, it was really important
to do the GRADE methodology. And Jacques’ TRIPICU study certainly fueled
that, the strong recommendation, moderate quality pediatric evidence, which had a 1B
rating. Five of our recommendations were strong recommendations
but had weak quality pediatric evidence, and 11 of them were weak recommendations with
low-quality pediatric evidence. So there is a little bit of evidence, but
not as robust as the adult world. And then 36 were, again, consensus panel expertise. And 45 of the recommendations were research
recommendations, really focusing those research priorities and where do we go from here. So the decision tree that Stacy had mentioned
already is part of the main article that we tried as best as possible to summarize the
data that we had put together and recommendation form into a decision tree that could be easy
to use and actually provide specific guidance to pediatric critical care practitioners around
the world. And we’re going to go through each section
a little bit just to give you a sense of what specific recommendations are. And then in this decision tree, we also tried
to incorporate as much of the recommendations, also the strength of evidence. You’ll find A through D throughout the recommendations. So really, embedded a lot of the process and
a lot of the recommendations within this one decision tree to help guide decision making
surrounding transfusion. Before getting to the decision tree, we actually
had some general statements that we felt, as a group, we wanted to put forth. They’re not necessarily recommendations, but
general ways of approaching transfusions. We felt that when there is a decision to transfuse,
we wanted to make it very clear that hemoglobin alone is not the only thing to consider. There’s an overall clinical context– signs,
symptoms, physiologic markers of oxygen debt– that should weigh into the decision as well
as the risks and benefits and alternatives to transfusion. And this we call the Good Practice Statement. And it’s really just a sounding board before
going into any decision making in transfusion. We also wanted to make sure that there’s a
clear statement that anemia should be appropriately managed, investigated, and that transfusion
in and of itself isn’t the answer, but really, causes of anemia need to be addressed. And then also, adoption of patient blood management
principles should be implemented to try to help avoid anemia. So the first branch point of our decision
tree was that we started with a critically ill child or a child at risk for critical
illness. That was our definition of our PICU population. Easily, the first branch is if you’re in hemorrhagic
shock. There’s really no controversy about giving
red blood transfusion if someone has hemorrhagic shock. But we did want to make a recommendation that
there is guidelines out there, and there’s expert consensus panel recommendation to use
a red blood cell plasma platelet ratio of two red blood cells to one plasma to one platelet
or a 1:1:1 when you have a hemorrhagic shock that was– you use that ratio until you no
longer have life-threatening bleeding. The next branch point if you’re not in hemorrhagic
shock is looking at your hemoglobin value, so if your hemoglobin is less than 5 versus
5 to 7. In a “less than 5”, we made the recommendation
to transfuse red cell. There is some– not high-quality, but some
evidence that hemoglobin less than 5 is associated with increasing morbidity and mortality. In critically ill children with a hemoglobin
between 5 to 7, there’s insufficient evidence. We really don’t know what to do here. And in that, we discussed that it’s reasonable
to consider transfusion, but based on clinical judgment in these children. There is discussion of ongoing studies to
consider using 6 versus 7 as a transfusion threshold, and so that’s in the research world. But there’s insufficient evidence to make
a specific recommendation between 5 to 7. We just don’t know. And so that gets back to the good practice
statement of clinical judgment included not just the hemoglobin level itself, but what
other factors with the patient should drive you to transfusion. And if you don’t have anything else besides
hemoglobin, we’re not sure that transfusion is really indicated in that situation. So that’s where clinical judgment– is how
we defined it. So if you actually had a hemoglobin greater
than 7, we had different branch points for those patients, either hemodynamically stable
or hemodynamically unstable. So the hemodynamically unstable, so patients
who are coming in septic shock who are still on active resuscitation phase: there really
was, unfortunately, a paucity of literature to allow us to make specific recommendations
about this. So we made a recommendation that we cannot
make a recommendation in that we feel that those patients are going to need their own
study. That was part of our knowledge gap, so that
is an area of ongoing need for us to make further recommendations. If you’re hemodynamically stable and hemoglobin
is greater than 7, that was an area that we spent the majority of our recommendations
on, are the different subgroups. We were able to look at if you did have hemodynamic
stability and you had the different criteria, then we made very clear recommendations about
not transfusing. And we had different subgroups: general critically
ill child, post-operative or surgical patients, respiratory failure excluding severe pediatric
ARDS, sepsis who are hemodynamically stable, non-life-threatening bleeding, or anybody
requiring renal replacement therapy. Any patient that had any of those criteria,
we made a strong recommendation not to transfuse those patients if their hemoglobin was greater
than 7. In the acute brain injury patient, we made
the recommendation to consider red cell transfusions, because we don’t have enough strong evidence
to say not to transfuse. So the recommendations in those patients were
to consider transfusion if the hemoglobin is between 7 to 10, and that further research
was needed. Stacey, before we leave acute brain injury,
were there representatives from the Brain Trauma Foundation guidelines on your development
committee so that there was a coordination of the assessment of the evidence? Yes, there was. Dr. Robert Tasker was on T.A.X.I., one of
our experts. We’re very lucky to have him. And he provided a tremendous amount of support
looking at recommendations for acute brain injury. And the situation for stem cell transplant
patients? So that, we said to consider red cell transfusions
between a hemoglobin of 7 to 8. And that was really driven by the data in
oncology. A lot of the data had 8 as a threshold instead
of what– a lot of the research in the critical care world had 7. So we had to say between 7 and 8, which is
where that came from. We had other more systemic illnesses, like
allo- or autoimmune hemolytic anemia, severe PARDS, defined as based on your oxygenation
index, or patients on ECMO or other supportive devices like VAD, for who we used the “use
clinical judgment” as our recommendation, kind of the same thing I said before, that
it wasn’t red cell alone, but using clinical judgment as a guide for doing red blood cells. That’s a wonderful overview and terrific guidance
for so many disorders in the context of hemodynamic stability. But I can’t resist asking you both, as experts
who are experts in the research and who care for critically ill children; I’m struggling
with how are we ever going to do a study in the hemodynamic instability context, which,
as you noted, Scot, you couldn’t reach recommendations on because there is such little literature? But could you see yourself randomizing a patient
with hemodynamic instability because they had severe pediatric ARDS or they were on
ECMO or threatening the need for ECMO? Do you feel like you could be at equipoise
to do such a study? I feel like the hemodynamic unstable patient
was an area that we really wanted as a group to be able to make recommendations on, and
we were unable to because it really wasn’t the data. So it became a very large knowledge gap to
us. But what was truly fascinating as part of
this process is to see how those knowledge gaps are galvanizing our T.A.X.I. group. So there is actually– a few of our experts
are actually involved in spearheading ongoing research, including a randomized controlled
trial in ECMO patients, a randomized trial in hemodynamically unstable patients, specifically
to answer these questions that we don’t know. And so it’s really– I think that there is
enough data to suggest that maybe transfusions may not be helpful in critical illness, that
the– I think that the equipoise has been generated to allow us to do this research. And I think that we really needed to come
together and recognize that there are risks to transfusion, and that we really should
try to avoid unnecessary transfusions. And we don’t know, in those hemodynamically
unstable patients, whether red blood cells are the right therapy or not. It could be other volume should be– would
be better. We don’t know. And I think that’s really– that unknown was
very clear in the room on many times. And so that has spurred a lot of interest
in pushing the envelope of what we can do. Well, that’s a terrific overview of so many
of the common disorders in the care of a critically ill child. But of course, there is one very large domain
that we haven’t discussed yet, and that is what is the evidence in the care of a child
with congenital heart disease. Do we have evidence to make recommendations? Excellent question. We do have preliminary– we do have evidence
based on small studies, actually small RCT studies performed by Jill Cholette, who was
fortunate to be on our T.A.X.I. group, looking at transfusion thresholds in congenital heart
disease. And in these studies, we looked at– we broke
it down into uncorrected congenital heart disease, biventricular repair, single ventricles
in stage one palliation, 2 or 3, as well as congenital-acquired myocardial dysfunction
and pulmonary hypertension. So in these preliminary studies, what we’re
seeing is that there is evidence that we can lower that threshold, which is also a very
big jump in this field and in the congenital heart disease patient. With uncorrected congenital heart disease,
our recommendations were to transfuse red cells to maintain a hemoglobin of 7 to 9 depending
on their cardiac reserve, which is a shift, certainly, down. Biventricular repair, the recommendation was
not to give a transfusion if the hemoglobin was above 7. And the single ventricle stage 1 palliation,
no transfusion of the hemoglobin was greater than 9, which is also a shift. And there’s adequate oxygenation and adequate
and normal end organ function, so we do have to remember those good practice statements
when we think about transfusion. But if there are those, if they do have adequate
oxygenation and normal end organ function, we don’t have to transfuse if the hemoglobin
is over 9. And the single ventricle stage 2 or 3, the
same– no transfusion of the hemoglobin is over 9. And the congenital or acquired myocardial
dysfunction, there is insufficient evidence, as well as the pulmonary hypertension world–
insufficient evidence. But what we did say is that there is no evidence
that transfusion above 10 is beneficial. We just don’t know how low to go. And so this was really spearheading that population
in transfusion practices. A couple of the other recommendations we had
about specific populations: sickle cell disease: so the requirement for a recommendation for
exchange transfusion if you have unstable acute chest syndrome or a simple transfusion
if you have compensated acute chest syndrome. And also, if a sickle cell patient is undergoing
a simple procedure or a surgical procedure with required anesthesia, we recommend that
the target hemoglobin of 10 is reached to decrease the amount of hemoglobin S. The last
recommendations are alternative processing for blood products. So we had blood bankers on the T.A.X.I. initiative
as well, and they recommended the two things we asked for: irradiation and washing. Irradiation was recommended for red blood
cells if there is a child at risk for transfusion-associated graft versus host disease or due to severe
congenital or acquired causes of immune deficiency in that specific population. And the washed blood cells were only if you
had a history of severe reaction to red blood cells in the past. If you’re caring for a hemodynamically unstable
patient tomorrow, and it’s a patient– it’s a child of six years old who’s been previously
well, but they’re in septic shock, and they have multiple organ failure that’s evolving. And you don’t know the trajectory of this. You’re wondering, is this patient going to
need ECMO? Is this patient going to survive? This looks very serious. And the patient’s blood pressure is low, and
you’ve added vasoactive infusions. And now the question becomes, should you increase
oxygen delivery? What are you going to do? What hemoglobin threshold are you going to
do in that patient? I think that you’re asking– you’re exposing
the hole in the recommendations because we weren’t able to make recommendations on that
specific population. I think that if we look back and say, OK,
what percentage of transfusions in the PICU are actually going to fit into that category,
they’re probably a lot less than you think. The majority of the transfusions that we found
are actually in populations that we did try to provide guidance on. And I think that what we do in our institution
is if anybody is being transfused and if they’re not following criteria, we actually have two
people who are going to make you very accountable to it very quickly. And I think that having a champion at your
institution, or having what Stacey mentioned before– patient blood management strategies
to provide guidelines for transfusion across the board– I think is really critical. When we actually get to an area that we don’t
have recommendations– say it was an ECMO patient or someone who’s unstable, like you
mentioned– we really don’t know enough to say this hemoglobin threshold should be used. But I feel like we have at least provided
a framework to say how do you put in clinical judgment so that you can actually use things
like hemoglobin as one factor. What other factors are suggesting that there’s
an oxygen deficit? We also need to recognize that red blood cells
may actually not improve oxygen delivery in a truly– true massive inflammatory state,
that the foreign red cells may actually not provide adequate oxygen that we need. So I think that we fall back to what is it
that we feel will– taking all those factors together will make a transfusion threshold
more amenable to giving a transfusion? So we will still give transfusions for unstable
patients, probably, if their hemoglobin is 9 or 10, even if there is a sense that there
is a poor oxygen deficit in that resuscitation phase. I feel like the resuscitation phase– there’s
still a sense that we should be giving transfusions more than we would do as stable patients. But I think that resuscitation phase is actually
hopefully relatively short in the grand scheme of that patient. And once you’re actually into the stabilization
phase, then we have very clear guidance on what to do. So in that patient that I described, you would
transfuse if their hemoglobin was 7, and their lactate is just a little bit elevated, and
their mixed venous is just a little bit low? I wish we had a true physiologic marker that
could really– besides hemoglobin– that says, yes, this is a transfusable patient. But really, the data is not clear enough to
say that there’s one marker of that. So you have to take the whole patient as a
whole and decide whether that patient could potentially benefit during the resuscitation
phase of the illness to have a transfusion, and knowing that you’re potentially putting
them at risk with inflammatory complications related to foreign red blood cells. So Scott, Stacey, that’s just a wonderful
overview and really helpful guidance. I’m sitting here thinking this is just so
helpful in terms of decision making at the bedside, which was your goal. But now where do we go next? That’s a great question. And we want to ensure the dissemination of
the T.A.X.I. guidelines as well as implementation. So currently, we’re actually studying the
implementation of the T.A.X.I. guidelines. Those implementations scientists that were
part of the T.A.X.I. process have developed a multi-center randomized control trial. They have eight centers already, randomizing
four to four, and are looking at using implementation strategies– the decision tree, clinical decision
support, a multi-modal, multidisciplinary strategy of physician behaviors– to understand
how T.A.X.I. is implemented at the bedside, and does it improve outcomes. We’re valuing the feasibility, acceptability,
and impact on practice. So we’re very excited that this is currently
underway. And we also thought that was very important
when we started T.A.X.I. to mention that we wanted to see– not only did we put these
out there, but we wanted to study it post. And so that was a decision made very early
on. So we’re very excited that this is underway. The next part is to conduct the next phases
of TAXI. So at the very beginning, we talked about
the first series was on red cell transfusions. So we are currently underway. And the second form of T.A.X.I., named T.A.X.I.-C.A.B.
is control and avoidance of bleeding. And this is to develop comprehensive, evidence-based,
and when evidence is lacking, expert-based recommendations on plasma and platelet use
in critically ill children. This is led by Marianne Nellis, and it focuses
on indication thresholds for prophylaxis and therapeutic transfusions. These are the subgroups of looking at plasma
and platelet use. And the final is to update the recommendations
periodically. So this is not a static process, and we recognize
that new literature is continually coming out on transfusion. And so the plan is, every five to 10 years,
to update the entire T.A.X.I. process and to really provide that update similar to other
guidelines that we have for critically ill children, to provide that update on T.A.X.I.
to make it a living guideline. We want to spur new research. Clearly, our T.A.X.I. experts are involved
with some new initiatives out there, randomized controlled trials looking at either lowering
the hemoglobin threshold even further, use in ECMO, use in hemodynamically unstable patients. There’s a lot of activity that’s been spurred
by this. And I think we’re hoping that there’s data
that’s going to make these recommendations more robust, more guideline-focused. And I think that in five years, at the very
least, we’re going to look again. Maybe it’s already been a year since publication,
and we already know there’s some data that we’re going to put into our recommendations. So it’s an exciting time for how these recommendations
will evolve. Well, Dr. Scot Bateman and Dr. Stacy Valentine
from the University of Massachusetts Medical School, thank you for sharing your expertise
today and developing a very helpful process and now recommendations for colleagues around
the world to improve care for the critically ill child. Thanks for having us. Thank you.

2 Replies to ““The Pediatric Critical Care Transfusion and Anemia Expertise Initiative” S. Bateman, S. Valentine”

  1. I am very grateful for your contribution to pediatrics training, especially important for distance learning and greatly appreciate the professionalism and talent in teaching material and disseminating your experience. Thank You, OPENPEDIATRICS

Leave a Reply

Your email address will not be published. Required fields are marked *