All posts by TheTraumaPro

Best Of EAST #4: 4-Factor PCC vs Andexanet Alfa For Factor Xa Inhibitor Reversal

Falls are by far the most common mechanism of injuries in US trauma centers these days. They typically occur in elderly patients, and a growing number are on some type of oral anticoagulant for their medical conditions. And the number of these patients who are taking a DOAC (direct thrombin inhibitor or factor Xa antagonist) is rising quickly.

Unfortunately, most of the DOACs do not have good reversal agents, and they are very, very expensive. Specifically, Andexanet Alfa, the antidote for rivaroxaban and apixaban used to cost in excess of $50,000 per dose. This has come down over time to “only” $22,000 per dose. Unfortunately, the half-life is much shorter than the agent it is neutralizing, frequently requiring two doses. And the kicker is that there are no studies definitively showing that Andexanet Alfa improves mortality when used for CNS hemorrhage.

Prothrombin complex concentrate (PCC) has been used for reversal of these agents as well. Its efficacy is also not well known. The group at George Washington University is presenting an abstract comparing it against Andexanet Alfa (AA) for reversal of either of the Factor Xa inhibitors (rivaroxaban, apixaban). They performed a multicenter study involving 10 trauma centers. The endpoints studied were number of transfusions, mortality, and ICU length of stay.

Here are the factoids:

  • From a total of 263 patients, 77 received AA and 186 received PCC
  • Only 4% of patients received a second dose of AA despite its short half-life
  • There was no significant difference in the number of PRBCs transfused
  • The authors stated that the mortality was significantly lower with PCC but the p value in the data table provided was = 0.05
  • They also stated that the ICU LOS was significantly lower with PCC (1.2 vs 1.5 days, p = 0.04)

The authors concluded that PCC is non-inferior to AA for reversal in bleeding trauma patients. They recommended a randomized study be done.

Bottom line: The first thing for you to know is that I have never been impressed with the data on Andexanet Alfa. Which means I have to be very careful and aware of my own cognitive bias. In practice, this means I can’t just look at the study title or abstract and be happy that it meets my confirmation bias. I have to make a conscious effort to critically read the paper or abstract and see if it really does mean what I want it to mean, or if I need to change my opinion.

This abstract doesn’t really satisfy my confirmation bias. The title states that PCC is not inferior to AA. I would certainly like to believe that. But in order to safely say that, it is vitally important that a power analysis is performed to ensure that enough patients are present in both treatment groups to confidently state that there was no difference. If the number of patients is too small, significance can’t be detected and non-inferiority cannot be confirmed.

The body of the abstract claims that mortality was significantly lower in the PCC group, although the table states that the p value was 0.05, which technically is not significant. The difference in mortality numbers is impressive (PCC mortality 20% vs 32% for AA) so why the significance issue?

And one note about significance. Be careful not to conflate statistical significance with real-life significance. ICU length of stay in this study was statistically significantly shorter in the PCC group (1.2 vs 1.5 days) but I doubt that a difference of 7 hours in the ICU is clinically relevant.

Here are my questions for the authors and presenter:

  • Did you have enough patients in the study to assure that the PCC treatment was actually non-inferior? Please show us your power analysis.
  • What were the inclusion criteria for the study? This will help us understand the patient group better. Were these primarily head bleeds, actual external or intra-cavity hemorrhage?
  • Please clarify the significance claim for mortality. The raw percentages are impressively different, but the P value is not significant.
  • Could the low rate of administering a second dose of AA have influenced the outcomes? As mentioned above, the half-life of the antidote is much shorter than that of the DOAC. Perhaps giving a second dose is actually needed and could have moved the results in favor of AA.

This is a thought-provoking abstract for me. Let’s see if you can either confirm or refute my opinion on AA!

Reference: 4-FACTOR PROTHROMBIN COMPLEX CONCENTRATE IS NOT INFERIOR TO ANDEXANET ALFA FOR THE REVERSAL OF FACTOR XA INHIBITORS: AN EAST MULTICENTER STUDY. EAST 25th ASA, oral abstract #15.

Best Of EAST #3: More Blood Product Stuff! Cryo vs Fibrinogen Concentrate

I’ve got more blood transfusion stuff for you today! This one isn’t exactly MTP, but it does deal with patients suffering from blood loss. Major trauma patients requiring transfusion are losing all of their clotting factors in addition to red cells and platelets. Some factors can be replaced by giving plasma. However, clotting factor I (fibrinogen) is not one of them. For that reason, most MTPs include administration of cryoprecipitate to replace it at key points in the algorithm.

However, fibrinogen replacement actually is available in two forms, cryoprecipitate (cryo) and fibrinogen concentrate (FC). These two forms have very different properties and costs. The authors from the University of Arizona Tucson massaged two years worth of data from the TQIP database. They included all trauma patients who received either cryo or FC, except for those with known clotting issues from bleeding disorders, liver disease, or anticoagulants they were taking. They analyzed differences in mortality, complications, transfusion requirements, and length of stay.

Here are the factoids:

  • A total of 85 patients received FC and 170 received cryo (?)
  • Blood product usage (red cells, plasma, and platelets) was less in the fibrinogen concentrate group (see table)
  • ICU and hospital lengths of stay were less in the FC group (see table)
  • There was no difference in complications between the two groups

The authors concluded that use of fibrinogen concentrate was associated with improvement in the outcomes they measured. They recommended that further studies be done to evaluate use of FC.

Bottom line: This study suggests that use of fibrinogen concentrate may be better than cryoprecipitate. But there are a lot of things to think about before jumping to conclusions. First, remember that this is an association paper. It cannot detect cause and effect. Blood component usage and LOS were indeed less in the FC group. But how many hundreds of other variables factor into those outcomes: specific injury pattern, ISS, body temperature, resuscitation balance, time to OR, and product availability are just a few. These and a host of others cannot be controlled easily in a study using TQIP data alone.

Next, look at the number of units of blood products given as listed in the table above. These are the average number of units given within 24 hours! These are relatively low numbers, and below what most would consider a “massive transfusion” threshold. Most of the time, we wouldn’t even think to measure fibrinogen in patients with such low blood replacement needs. Perhaps we should?

And finally, I have questions about the low number of subjects in the study. There are only 255 patients analyzed, and propensity score matching was used. This implies that overall numbers were low and that several covariates had to be considered. It seems like more subjects should have been available in the tens of thousands of patients submitted in two years to TQIP

Here are my questions for the authors and presenters:

  • Please review your statistical analysis with the audience. Why is the number of subjects so low? What did you control for in your propensity score matching and how did this impact your n?
  • Comment on the low numbers of blood products given. Do you think that patients who have such a relatively low transfusion requirement even needed a fibrinogen supplement?
  • Did you consider performing this analysis on patients who underwent massive transfusion? The effect might have been more pronounced in that group.
  • Since fibrinogen concentrate costs about 3x what cryo does, how would this factor into the math involved in deciding to use FC?

This study raises a lot of interesting questions. I hope the authors incorporate the answers into their presentation!

Reference: FIBRINOGEN SUPPLEMENTATION FOR TRAUMA PATIENTS: SHOULD YOU CHOOSE FIBRINOGEN CONCENTRATE OVER CRYOPRECIPITATE? EAST 25th ASA, oral abstract #3.

Best Of EAST #2: Pay Attention To Platelet Ratios In Your MTP!

More MTP stuff! Every trauma center has a massive transfusion protocol, and current literature encourages them to try to achieve an “optimal” transfusion ratio. The literature has converged on a red cell to plasma ratio of somewhere between 1:1 and 2:1. Less has been written about platelet ratios, and trauma centers often don’t pay as much attention to this ratio when reviewing MTPs.

But is it important? The trauma group at the Massachusetts General Hospital examined the impact of platelet ratios on mortality in patients undergoing MTP. This was another TQIP data analysis, performed over a nine year period.

The authors defined massive transfusion as ten or more units of PRBC in the first 24 hours, or any number of units of red cells, plasma, or platelets given within the first four hours. They also defined “balanced” as a ratio of RBC to FFP and RBC to platelets <2. Multivariate regression analysis was performed to gauge the impact of ratios and achievement of a balanced resuscitation on 24-hour mortality.

Here are the factoids:

  • A total of 7,520 patients in the dataset underwent MTP
  • Nearly 83% achieved RBC to FFP balance, but only 6% had RBC to platelet balance (!)
  • Patients with both balanced FFP and platelets had the lowest mortality at 24 hours
  • Mortality increased by 2x with unbalanced plasma, a little more than 2x with unbalanced platelets, and 3x if both were out of balance (see figure)

The authors concluded that the platelet component of the MTP was frequently out of balance, and that it is associated with mortality to a greater degree than with unbalanced plasma.

Bottom line: This paper confirms my observations that trauma centers pay a lot more attention to the red cell to plasma ratio and don’t get as excited when the platelets are out of line. Part of this is probably due to confusion over how to count platelet packs. Typically they are delivered in packs called “pheresis” or “apheresis.” Each is the equivalent of about 6 units of platelets (check with your blood bank for more exact numbers). This means that a ratio of 6 RBC to 5 plasma to 1 platelets would be considered balanced. But a ratio of 28:28:2 would not.

According to this abstract, the use of sufficient platelets is important. This makes sense. However, the exact mechanism cannot be determined from this type of study. It could be a direct effect of not having enough platelets to form good clot. Or it could be something completely outside the clotting mechanism, just an association with something in the care processes that occurs as these patients undergo resuscitation. 

The why doesn’t matter so much, though. This abstract presents compelling data that suggests that we really need to pay attention to the platelet ratios given during the MTP. They should be analyzed just as closely as plasma ratios during PI review, and changes to the MTP process implemented to normalize this important ratio.

Here are my questions for the authors and presenter:

  • There is a statement in the methods section that is not clear. “only patients with steady RBC/PLT and RBC/FFP ratios between 4-and 24-hr were analyzed.” What is your definition of “steady?”
  • Did you see any mortality patterns in the data you analyzed that might suggest why lower platelet volumes were more deadly?

This was a nicely done abstract, and I look forward to the live presentation and the finished manuscript!

Reference: DON’T FORGET THE PLATELETS: BALANCED TRANSFUSION AND THE INDEPENDENT IMPACT OF RBC/PLT RATIO ON MORTALITY IN MASSIVELY TRANSFUSED TRAUMA PATIENTS. EAST 25th ASA, Oral abstract #1.

Best Of EAST #1: When Is MTP Blood Use Too Much?

The 35th Annual Scientific Assembly of the Eastern Association for the Surgery of Trauma (EAST) begins in only a month! I will be there, sitting in the front row listening to all the great presentations. As usual, I have selected some of the abstracts that I find most interesting and will be sharing my thoughts on them with you over the coming weeks.

Let’s start out with a paper about the massive transfusion protocol (MTP). Blood has always been a scarce resource. And now, thanks to COVID, it is becoming even more so. Every trauma professional reading this has likely been involved in a trauma resuscitation that has used dozens of units of blood and other products. Unfortunately, most of the patients who require this much do not survive.

How does one balance the rapid use of many, many units of blood products with the (un)likelihood of survival and the impact of having less blood for other patients in your hospital or future incoming trauma patients? In other words, when does the use of additional blood become futile? Until now, there have been no real answers to these questions.

The trauma group at George Washington University did a deep dive into the TQIP database seeking some guidance on this topic. They reviewed five years of data, targeting patients who received at least one unit of blood within four hours of arrival. Four-hour and 24-hour mortality was analyzed to determine the point at which additional blood products did not improve survival.

The authors looked at the data two ways. They analyzed the results for all comers, as well as for patients who received balanced resuscitation. Balanced was defined as a red cell to plasma ratio in the range of 1:1 to 2:1. Results were controlled as best as possible for age, sex, race, highest AIS in each body region, comorbidities, advanced directives, and the type of surgery performed to control bleeding.

Here are the factoids:

  • Nearly 100,000 patient records were analyzed, and about 30,000 patients were found to have balanced resuscitation
  • In the all-comers group, mortality plateaued after 41 units at 4 hours and 53 units at 24 hours
  • In the balanced resuscitation patients, mortality plateaued at 40 units (4 hours) and 41 units (24 hours)

The authors concluded that this data should be used as markers for resuscitative timeouts to assess the plan of care.

My comments: This paper is very focused and provides some apparently straightforward results. However, it required some sophisticated statistical analysis to sift through the many variables that need to be controlled to obtain meaningful results. From reading the abstract, it appears that they did a good job of this.

I believe the lower number of units needed by 24 hours in the balanced resuscitation group demonstrates the benefit of getting the MTP ratios right. Non-balanced resuscitation is less efficient / effective and requires the use of more products to hit the mortality plateau.

This paper supports my opinion that a resuscitation timeout is a useful tool in helping us protect our valuable blood product resources and ensuring availability for as many patients in need as possible. What would this look like? Here are my thoughts:

  • Assign one person to monitor the MTP process in real-time. This obviously cannot be the surgeon or a member of the anesthesia team. Or even the operating room crew, as everyone will be very busy. The best practice I’ve seen is to have a dedicated trauma nurse or APP in the ED/OR recording the process on a specialized form and directing which units to give to keep the resuscitation balanced.
  • Call a timeout when the magic threshold is reached. This paper suggests that 40 is a good number.
  • Require that another trauma surgeon come into the room and review the patient condition, operative findings, and progress thus far. The two surgeons should then come to a consensus regarding utility vs futility of further surgery. Based on that decision, the operative procedure either continues or stops.
  • If the operation is to continue, then more timeouts should occur after a defined number of additional products are given.

Here are my questions for the authors / presenter:

  • The statistical analysis required is fairly advanced. Please explain in simple language why the specific regression analysis with bootstrapping was selected.
  • How do you envision applying the thresholds discovered in your paper?

This is an exciting paper and provides important information about the MTP process. I’m looking forward to hearing it in person!

Reference: CRESTING MORTALITY: DEFINING A PLATEAU IN ONGOING MASSIVE TRANSFUSION, EAST 25th ASA, oral abstract #14.

Retained Hemothorax: The Practice Guideline

Over the last few days, I’ve reviewed some data on managing hemothorax, as well as the use of lytics. Then I looked at a paper describing one institution’s experience dealing with retained hemothorax, including the use of VATS. But there really isn’t much out there on how to roll all this together.

Until now. The trauma group at Vanderbilt published a paper describing their experience with a home-grown practice guideline for managing retained hemothorax.  Here’s what it looks like:

I know it’s small, so just click it to download a pdf copy. I’ve simplified the flow a little as well.

All stable patients with hemothorax admitted to the trauma service were included over a 2.5 year period. The practice guideline was implemented midway through this study period. Before implementation, patients were treated at the discretion of the surgeon. Afterwards, the practice guideline was followed.

Here are the factoids:

  • There were an equal number of patients pre- and post-guideline implementation (326 vs 316)
  • An equal proportion of each group required an initial intervention, generally a chest tube (69% vs 65%)
  • The number of patients requiring an additional intervention (chest tube, VATS, lytics, etc) decreased significantly from 15% to 9%
  • Empyema rate was unchanged at 2.5%
  • Use of VATS decreased significantly from 8% to 3%
  • Use of catheter guided drainage increased significantly from 0.6% to 3%
  • Hospital length of stay was the same, ranging from 4 to 11 days (much shorter than the lytics studies!)

Bottom line: This is how design of practice guidelines is supposed to work. Identify a problem, typically a clinical issue with a large amount of provider care variability. Look at the literature. In general, find it of little help. Design a practical guideline that covers the major issues. Implement, monitor, and analyze. Tweak as necessary based on lessons learned. If you wait for the definitive study to guide you, you’ll be waiting for a long time.

This study did not significantly change outcomes like hospital stay or complications. But it did decrease the number of more invasive procedures and decreased variability of care, with the attendant benefits from both of these. It also dictates more selective (and intelligent) use of additional tubes, catheters, and lytics. 

I like this so much that I’ve incorporated parts of it into the chest tube guideline at my center!

Download the practice guideline here.

Related posts:

Reference: Use of an evidence-based algorithm for patients with traumatic hemothorax reduces need for additional interventions. J Trauma 82(4):728-732, 2017.