Category Archives: Abstracts

Best of EAST #8: Reversing Antithrombotic Drugs After Severe TBI

January 11, 2021 The Trauma Pro Leave a comment

Falls are the most common mechanism of injury at a majority of trauma centers these days. And due to the escalating number of comorbidities in our older population, more and more are taking some kind of anticoagulant or antiplatelet medication. And as all trauma professionals know, falling down and failure to clot do not mix well.

A variety of reversal regimens have been developed, including Vitamin K, plasma or platelet infusion, prothrombin complex concentrate, andexxanet, or idarucizumab depending on the agent. But when it comes to evaluating the efficacy of these agents, there are two important questions that need to be answered:

Does the regimen reverse or neutralize the offending agent?
and more importantly
Does the regimen have a positive effect, i.e. reduce mortality and/or complications?

This last question has been problematic, especially for the direct oral anticoagulant drugs (DOACs). They are very expensive, but there has been little, if any, evidence that they improve mortality.

A study from the University of Florida at Jacksonville, and sponsored by EAST was performed last year. It was a multi-center, prospective, observational study of data provided by 15 US trauma centers. They collected data on the agents used, reversal attempts, and comorbidities in injured patients taking these drugs, and analyzed for head injury severity and mortality.

Here are the factoids:

There were a total of 2913 patients in the study, 46% on aspirin (ASA), 13% taking ASA and a P2Y12 inhibitor (one of the -grels), 11% on warfarin, 4% on ASA + warfarin, 13.5% on a Factor Xa inhibitor, and 6% on a Xa inhibitor + ASA
Patients on platelet blockers (P2Y12 inhibitor) had the highest mean ISS at 9
Warfarin was associated with a higher abbreviated injury score (AIS) for head, 1.2
Controlling for ISS, comorbidities, ISS, and initial SBP, warfarin + ASA had the highest head ISS with an odds ratio of 2.1 (with the lower confidence interval value of 1.19)
Reversal of antiplatelet therapy with DDAVP was not successful, with no change in mortality (87% with reversal and 93% without)
Reversal of Xa inhibitors with plasma or PCC was also unsuccessful with a mortality of 100% with reversal and 95% without

The authors concluded that reversal attempts for antiplatelet therapy or Factor Xa inhibitors did not decrease mortality, and shared the observation that combination therapies posed the most risk for severity of head injury.

My comments: Remember, the first thing to do is look at the study group. The authors did not share the inclusion or exclusion criteria for the study in the abstract, so we are a little in the dark here.

The next item that makes this study difficult to interpret (and perform) is the fact that nearly a quarter are on combination therapy for their anticoagulation. So even though nearly 3,000 patients were studied, many of the medication subgroups had only a few hundred subjects. The aspirin group was the largest, with 1,338. This makes me wonder if the overall study had the statistical power to find subtle differences in their outcome measures and support the conclusions.

Now have a look at one of the results tables:

In reviewing the demographic data, the concept of statistical significance vs clinical significance quickly comes to mind. Somehow, age, ISS, head AIS, mortality, and SBP are significantly different between some of the groups. Yet if you examine the specific values across most of the rows, there is little difference (e.g SBP ranges from 137 to 147, ISS from 7-9, mortality from 2-7%). These are all clinically identical. The only row that means much to me is the top one telling how many patients are in a group.

Here are my questions for the authors and presenter:

Tell us about the study design, especially the inclusion and exclusion criteria. Were there any? How might this have influenced the study group?
Please comment on your perception of the statistical power of the study, especially with seven groups of patients, each with relatively small numbers.
Do you have information on the variety of reversal agents used? Were there any standards? Could this have contributed to the mortality in some of the groups?
Do you have any clinical recommendations based on your findings? If not, what is the next step in examining this group of patients?

My bottom line is that I’m not sure that this study has the power to show us any significant differences. And looking at the information table and logistic regression results (odds ratio confidence intervals close to 1), I’m not really able to learn anything new from it. I’m hoping to learn a lot from the live presentation!

Reference: EAST MCT: comparison of pre-injury antithrombotic use and reversal strategies among severe TBI patients. EAST 2021, Paper 19.

Abstracts, Resuscitation

Best of EAST #7: Whole Blood Plus 4-Factor Prothrombin Complex Concentrate

January 8, 2021 The Trauma Pro Leave a comment

In my last post, I went through some of the basics of whole blood transfusion. However, the focus was more on compatibility than function. Today, I’ll review an abstract that explored functionality of that blood transfused.

In theory, whole blood contains the usual array of clotting factors. It has been shown that high factor levels persist in whole blood, even when stored at room temperature. So in theory, additional clotting factor infusion should not be necessary.

The group at the University of Arizona explored adding 4-factor prothrombin complex concentrate (4-PCC) to whole blood transfusion. The scanned three years of data in the TQIP database. They identified two groups of patients, those who received whole blood alone and those who received 4-PCC in addition to it. They were interested in the impact on total product transfused and the usual crude outcomes of hospital / ICU length of stay and mortality.

Here are the factoids:

Only 252 patients in this entire database (tens of thousands of records in three years) received whole blood, and 84 of them also received 4-PCC
The patients tended to be young (average age 47), 63% male, with moderate (median ISS 27), and blunt injury in 85%
Administration of 4-PCC was associated with a significantly decreased transfusion requirement of both blood (5 vs 8 units) and plasma (3 vs 6 units), but not platelets
ICU LOS was significantly lower in the 4-PCC group (5 vs 8 days), but there was no difference in hospital stay or in-hospital mortality

The authors concluded that 4-PCC given with whole blood was associated with a decrease in transfusion requirements and ICU length of stay, and that further studies were needed.

My comments: Well, this is certainly interesting and unexpected. Why would a clinician even think of giving 4-PCC when giving whole blood? It looks like a very rare occurrence in the dataset. Unfortunately, we can never find out. We can’t just go back and look in the charts. Perhaps these centers were using TEG or ROTEM during the resuscitation?

As always in these big databank analyses, the researchers can only control for the variables they can think of that are already present in the database. Although they were able to match the patient groups for the usual demographics, vital signs, injury patterns, comorbidities, and trauma center level, it is entirely possible that there were other factors in play.

Here are some questions for the authors and presenter:

Why did you choose to do this study? Was there some clinical question that arose that triggered it? Something you found in the literature that suggested it?
How do you explain the results, given that the factors in 4-PCC have been shown to persist at functional levels in whole blood? Why do you think less blood and plasma were needed?
What needs to happen next? I agree that more research is needed to see if this association is real. How would you go about doing it?

Thanks for a very intriguing paper! Details will follow, I’m sure.

Reference: Four factor prothrombin complex concentrate in adjunct to whole blood in trauma-related hemorrhage: does whole blood replace the need of factors? EAST 2021, Paper 18.

Abstracts, Resuscitation

Best of EAST #6: Does Rh Status Matter In Whole Blood Transfusion?

January 7, 2021 The Trauma Pro Leave a comment

What goes around comes around. Fifty plus years ago, the only transfusion product available was whole blood. Then the major blood banks discovered that more patients could be treated for specific problems if the blood were fractionated. Packed red cells then became the standard for trauma transfusion and persists to this day.

But there is a move afoot to re-explore the use of whole blood. There are many theoretical advantages, since our trauma patients are bleeding whole blood, not packed cells. Unfortunately, combining a unit of packed red cells, plasma, and platelets does not give you a reconstituted unit of whole blood by a long shot. Check out this diagram:

The challenge is that we are used to only thinking about universal donor red cells (group O Rh-). This is the safest packed cell product to give a patient with an unknown blood type. But unfortunately, it is also one of the hardest to find, present in about 7% of the population.

Packed red cells are nearly plasma free. What we don’t think about with whole blood is the level of antibodies to blood groups that are present in the plasma. Group O blood will have plasma with anti-A and anti-B antibodies. So if we include the plasma with those universal donor red cells, these antibodies may attack the patient’s red cells if he or she is group A, B, or AB and cause a reaction.

Theoretically, this issue can be avoided by using universal donor plasma (group AB+). Since the donor has all of the major group antigens, they will have no antibodies in their plasma. Unfortunately again, this is a rare type and tough to get donors (about 3% of the population).

To avoid potential transfusion reactions, group O whole blood is tested for antibody titers, and only low titer blood is selected for transfusion. Typically Rh- whole blood has been selected to avoid any issues with Rh incompatibility, even though reactions to this antigen are usually mild.

The group at the University of Texas – Houston reviewed their experience using Rh+ low titer group O blood in trauma resuscitations. Their two-year study substituted Rh+ whole blood when Rh- product was not available. They monitored patients for transfusion reactions, renal failure, sepsis, VTE, and ARDS.

Here are the factoids:

A total of 637 patients received low titer group O blood during the study period; 448 received Rh+ product and 189 received Rh-
Those receiving Rh+ blood were more likely to be male, had lower initial SBP, and a significantly lower GCS (7 vs 12)
Overall there were no differences in hemolysis labs, transfusion reaction, complications or mortality
The patient groups were then sliced and diced by their own Rh antibody status to see if Rh- patients had an increased likelihood of problems from Rh+ plasma
Once again, the Rh- subgroup was significantly different for sex (57% female vs 26% in the Rh+ group), and blunt trauma mechanism (92% vs 70%)
And once again no differences were seen in hemolysis, transfusion reaction, complications or mortality

The authors then concluded that Rh+ low titer whole blood is a safe alternative in either Rh+ or Rh- patients.

My comments: Sounds good, right? But wait a minute! This was a non-randomized observational study. It appears that Rh+ whole blood was used when Rh- was unavailable, which was quite a bit of the time. This is clear when you see the demographic differences listed above between the two recipient groups, as well as the subgroups stratified by their own Rh status.

This is the first thing that makes me a bit more skeptical of the recommendation. The other one is something you’ve heard me harp about before… non-inferiority studies. This abstract tries to say that since they did not detect a difference, then the two products are equivalent.

That is only true if there is adequate power in the number of patients studied. If not, you may not be able to show a statistically significant difference. By my own calculations, if the incidence of transfusion reaction in the Rh- group is 1% and the ratio of the patient groups is 0.42, the reported sample size could only show a significant difference if the Rh+ patients had a 5% transfusion reaction rate.

So is it truly non-inferior, or does the study need include a lot more patients?

Here are my questions for the authors and presenter:

What is the impact of the non-randomized patient selection process on your results? The groups and subgroups appear to be very different. Couldn’t this influence your results?
Exactly what type of statistical analysis did you use? Your abstract merely lists the software package, not the specific tests applied.
Do you believe that your study is sufficiently powered? What assumptions did you use to calculate this?

As we move toward more use of whole blood, the Rh question will be an important one. I look forward to questioning the authors on this one!

Reference: Can Rh+ whole blood be safely used as an alternative to Rh- product? An analysis of efforts to improve the sustainability of a hospital’s low titer group O whole blood program. EAST 2021, Paper 17.

Abstracts, Resuscitation

Best of EAST #1: Ultramassive Transfusion Survival

January 2, 2021 The Trauma Pro Leave a comment

All right, let’s kick of this EASTfest with an abstract from one of the Eastern Association for the Surgery of Trauma multicenter studies. This one looked at outcomes after what they term “ultra-massive” resuscitation.

There are a number of definitions for “massive transfusion” which I’ve discussed before. They are basically trauma resuscitations in which the massive transfusion protocol is triggered. The group that designed this study defined ultra-massive resuscitation as one that entails transfusing at least 20 units of packed red cells within 24 hours.

The study focused on factors predicting survival in these patients. They used multivariate logistic regression as well as another regression tool, classification and regression tree analysis (CART). They used these tools to control for age, ISS, mechanism of injury, base deficit, and crystalloid use.

Here are the factoids:

A total of 400 patients were studied at 15 trauma centers over an eleven year period
Subjects were young (mean 37 years), male (81%), severely injured (mean ISS 34) and in shock
Median transfused products were 29u PRBCc, 23u FFP, and 24u platelets
Mortality was high with half dying in 24 hours and two thirds not surviving to discharge
Transfusion ratios > 1.5:1 for both RBC to plasma and RBC to platelets were strongly association with death
CART identified severe head injury, resuscitative thoracotomy, and low platelet count (< 169K / microliter) we association with high mortality
The best chance for survival occurred in those without a head injury, no thoracotomy, and higher platelet count

The authors concluded that the failure to meet balanced resuscitation goals was the main concern for mortality, and recommended more attention to meeting ratios.

My comments: I’m not so sure I’ve learned a lot from this abstract. I think we already knew that people with severe TBI or thoracotomy don’t do very well, especially if they need that much blood.

I also worry about the heterogeneity of the population. The variables that were controlled still offer quite a bit of variability in the injuries and condition of these trauma patients. I think this will make it difficult to come to many solid conclusions when looking at something as crude as mortality.

Here are my questions for the authors and presenter:

Why are there so few patients? An eleven year study with 15 centers participating means that each submitted less than 3 cases per year. Most busy Level I centers have many more than that in a single year. Was there some other kind of data selection or limitation that is not described in the abstract? Do you think there is enough power? See question 3 for more on this.
How did you arrive at an admission platelet count threshold of 169,000/ul? This would seem to be a surrogate for something else going on, and I’m not sure what. But it just seems so arbitrary.
The transfusion ratios are a bit confusing. For ratios less than 1.5:1, there are no error bars. Does this mean that every one of those patients survived? That’s remarkable if so. And the error bars for the groups with a ratio > 1.5:1 are perilously close to the 1 line, and they have quite a range. Is the statistical power really there to convincingly show a difference? This is the most interesting part of the abstract, so please expound upon it.
Explain your use of CART. How did you determine the specific determine the specific thresholds used in the CART model? Why did you choose to use this tool? For my readers, here is the tree presented in the abstract.
What is the real message of the abstract? We already know that if patients who have a severe head injury or get their chest cracked are probably not going to make it. The transfusion ratio information is somewhat interesting, but there is better quality data out there that defines acceptable ratios. The platelet count information… interesting. What more do you have?

I think there is a lot of potential in this dataset once you overcome the small numbers. I’m very interested in the authors’ presentation!

Reference: Ultra-massive transfusion outcomes in a modern era: an EAST multicenter study. EAST 2021, Paper 1.

Abstracts

The Best Of EAST 2021

January 1, 2021 The Trauma Pro Leave a comment

The Eastern Association for the Surgery of Trauma annual meeting starts in just 2 weeks! Keeping to tradition, I’m going to start reviewing some of the more interesting (to me) abstracts to be presented at the meeting and sharing my thoughts with you.

There are 33 regular abstracts and 17 quick shot abstracts to be presented. I’m going to focus on the regular abstracts, since there will be an opportunity to question the authors (hopefully) at the virtual meeting. Quick shots are a very brief presentation only.

Let me share how I process a batch of abstracts like this. First, I capture the pdf file with all the abstracts and open it in a pdf markup program. Adobe Reader or Acrobat have basic capabilities, but I prefer a more full-featured product so I can scribble notes and stuff on it.

Now, I go through the file looking at titles. Keep in mind I am a clinical trauma surgeon. So right off the bat I will pretty much discard any bench type research. No matter how interesting it may sound, it will be years before it may (or more likely won’t) be clinically relevant. Invariably, I will pay no further attention to these.

If the title, suggests it is an animal study, I may consider it. But probably not. The research idea had better be a very interesting or intriguing one that should definitely stimulate further thought and research. If it’s just making an incremental advance, there won’t be any clinical relevance to humans for a few more years. There are some REBOA abstracts in the current batch that fall into this category. I do keep the research concept in my mind for future consideration when I see related papers, but for now I ignore.

Now, I am left with mostly clinically relevant papers. As I read the title I ask myself:

Did I know this already? If I did, I read the intro and conclusion to see if this abstract adds anything different to what I thought I knew. If it does, I’ll read the whole thing and analyze it. But most of the time, there is not enough novelty to keep me interested.
Is this truly something new and different? This is a very unusual occurrence. Most work adds incrementally to previous research. But if it really is new and different, I will latch onto this and read it in great detail.
Might it refine our approach to certain clinical problems? Could we improve the usual way we take care of our patients? These are of great interest to me. However, remember that no single paper (or certainly abstract) should ever make you change your practice. There are so many exciting things that have been published exactly once that don’t just pan out. Beware the one-hit wonder. And unfortunately, you don’t know it is one until months or years later when the concept has been disproven or no one else has been interested enough to duplicate it.
Have the authors used a new approach to tackle a problem? Exploring a new way to look at a specific problem may be generalized to other problems as well. So in this case I will forgive a boring or already known result so I can scrutinize a new research tool.

By now, I’ve cut the number of abstracts roughly in half. That’s still too many to write about. So finally, I have to narrow down the field by ranking in order of my interest level. I fully recognize that my interests will not be necessarily be perfectly aligned with yours. But I do know my audience, and most of you share the same areas of curiosity. Unfortunately, some good abstracts will be ignored. But there is one thing you can do: look over the abstract collection yourself and let me know about specific abstracts you would like to see discussed! I am happy to oblige.

So beginning tomorrow, I’ll post the most interesting EAST abstracts in program book order. I’ll provide the author’s description and my analysis. I will also list some questions that I (and probably you) have that the authors should consider. I always make a point of notifying the authors each day when I post about their abstract so they can study the questions and potentially address them in their virtual presentation.

And as always, if you have questions, suggestions, or abstracts you would like discussed, just reply here or on Twitter. I hope to “see” you at EAST!

Home of the Trauma Professional's Blog