Bleeding trauma patients present a couple of challenges to us in the Emergency Department. Increasingly, our aging population continues to engage in energetic daily pursuits while increasing numbers are prescribed newer direct oral anticoagulant treatments (we have a standing joke that the typical 80th birthday present is a stepladder and a box of apixaban), meaning that their trauma presentations are complicated by coagulopathy, which cannot be easily reversed.
In addition, researchers have written and spoken about a “coagulopathy of trauma“, something that definitely exists, is well-researched (particularly by people like Karim Brohi) but remains poorly understood. In short, treating these trauma patients is more tricky than simply “filling the bath” and “turning off the tap”.
In this recent, free and open access publication in Intensive Care Medicine, a group of authors have tried to ascertain whether point-of-care testing can direct the use of blood products and therapies to specifically address coagulation issues, as opposed to the rather blunt tool that is the major haemorrhage protocol. It’s a good question – since transfusion itself is not without risk to the patient, and blood itself is a scarce resource. Knowing whether we can treat each patient individually according to their needs makes logical and ethical sense, and the question being asked here is whether TEG and ROTEM can help us to do that.
Given that the paper is freely available, you might want to read it yourself before continuing to read here.
What is the paper about and what did they do?
The authors have set out to find out whether using point-of-care coagulation testing, referred to in the paper as viscoelastic haemostatic assays (VHA) to direct massive haemorrhage protocol use, as opposed to standard practice (conventional coagulation tests, CCT) would improve mortality and reduce the need for massive transfusion (10+ units of blood in the first 24h after injury). They looked at a number of secondary outcomes too, which we will dip into later, but it’s always helpful to focus on the primary outcome first as this is how they study is designed and powered (how the sample size is determined).
The study was carried out across seven major trauma centres in Europe, three of which were in the UK. It was a pragmatic trial, which makes sense when you start to think about how on earth you could recruit acutely injured patients to a trial like this without introducing lots of biases. Whether TEG or ROTEM was used was determined by each centre itself, according to availability of the test. The inclusion criteria were surprisingly broad, which is good as it potentially captures a realistic sample of trauma patients: adult patients, presenting with signs of bleeding activating the major haemorrhage protocol and with red cell transfusion commenced, were randomised within 3h of injury and no more than 1h after presentation to ED. Subjects were randomised in 1:1 ratio using block randomisation by centre.
Obviously, for treating clinicians to be able to use the results of the VHA to inform their decision-making as part of the study, they could not be blinded to the allocation of the patient, but as is good practice in these situations, the research team were blinded to the group allocation when they carried out their data analysis.
The patients received their local hospital’s major haemorrhage protocol as standard (1:1:1 ratio of red cells, plasma and platelets) and additional administration of therapies like platelets, fibrinogen, plasmin and antifibrinolytic agents was determined according to the results of the VHA or CCT. There is a separate TACTIC algorithm, published here, which defined threshold parameters and was used to guide the use of these therapies.
The authors decided that the best way to determine the effectiveness of using VHA as a guide to therapy was to look at the proportion of subjects alive and free of massive transfusion at 24h post injury. It’s worth just making sure here that we understand the distinction being made:
- all patients received the major haemorrhage protocol (MHP)
- the authors were interested in patients free from massive transfusion protocol (MTP); that is, administration of 10 or more units of packed red cells in the 24h since injury.
What’s interesting about the planning phase of this study was the relatively small sample size. The study was powered to detect a reduction in the primary outcome (patients dying or receiving MTP) from 28% (as seen in prior studies) to 15%, with a 5% significance level using two-sided tests. This gave a planned sample size of 392 in total, with 196 in each arm. This sounds pretty small to me, and while the maths makes sense I have alarm bells ringing when we are considering quite a heterogenous population across a number of centres. More on this later.
What did they find?
OK, remember the primary outcome estimate used to power the study? Well, in the standard care (CCT) group, 64% of subjects were alive and free of MTP and in the VHA group it was 67%. That implies that the rate of patients dying or receiving MTP (the proportion used in the sample size calculation) was 36%, which is quite different from the 28% used from other studies. This may affect the power of the study to detect a difference.
First of all, it’s helpful to look at Table 1 to see whether the patients were similar in both groups. There were similar proportions of blunt trauma (67% vs 66%), Injury Severity Score [ISS] (median 26, IQR 16-35 vs median 26, IQR 17-37), time from injury to admission (median 67min, IQR 42-98 vs median 70, IQR 48-95) and even the pre-baseline therapies (TXA, RBCs and other blood products transfused). This is quite impressive considering the small sample size and the fact that recruitment occurred across 7 different sites. What the paper doesn’t tell us is how recruitment was distributed across the centres; it’s entirely possible that the majority of patients were recruited at one or two centres with only a handful at each of the others. It’s worth considering how this might affect the wider applicability of results although of course this is speculation.
Thinking about this logically, if results are available to clinicians sooner, they are more likely to make decisions about therapy sooner, and that’s the first finding the authors describe. In the VHA group, patients received more additional interventions and received them sooner than those in the CCT group.
This, however, was not the primary outcome. Interestingly, there was no significant difference between the two groups in either mortality at 24h (17% in the CCT group, 14% in the VHA group; OR 0.83, 95% confidence interval 0.48-1.42) or in MTP use at 24h (28% in the CCT group, 26% in the VHA group; OR 0.91, 95% confidence interval 0.59-1.42). Remember that with odds ratios, we are looking for the confidence interval crossing 1.0 to determine whether the outcome might be equally likely in either group.
There are some interesting nuggets in the other data analyses too; the absence of significant difference in mortality between the two groups persisted in analysis up to 90 days. The amount of blood product received was similar between the patient groups, receiving a median of packed cells and four units of plasma.
When we have a look at Figure 2, we get an idea that there might have been some (non-statistically significant) differences between the two groups. The most notable is within the subgroup analysis of patients with severe traumatic brain injury, where 46% of those in the CCT group were alive and free of MTP at 24h, compared to 64% of those in the VHA group. Of course, this subgroup is small (74 in total) so a single patient either side could make a big difference, but it’s worth considering as an avenue for further research.
In addition, a greater proportion of patients in the CCT group who were taking anticoagulation prior to injury were alive and MTP free at 24h (67% vs 50%). This is really interesting, as our expectation would naturally be that targeted therapies in groups with pre-deranged (I’m not sure that’s a word but it should be) coagulation would be advantageous but instead we see the (implied, non-statistically significant) opposite. More food for thought. Remember, though, that although the OR does not cross 1.0 in these two subgroup analyses, they are small subgroups (74 patients with TBI, 27 patients with coagulopathy) and the study is not powered to detect significant differences in these outcomes.
What does the study mean for us?
The authors conclude that point-of-care testing identifies coagulation issues and directs therapy sooner than CCT alone, but crucially this does not translate to benefit to patients in the context of mortality or reduced requirement for blood products. It’s important that we know the answer to this question and while this study doesn’t quite reassure me that we have a definitive conclusion, I think it’s an excellent starting point.
On twitter, corresponding author and friend of St Emlyn’s Karim Brohi addressed some of the issues with the study itself.
In particular, he highlights how trauma coagulopathy itself is less prevalent than it was at the time the study was designed (2012), reflecting a likely change in trauma management practices and reduced use of crystalloids before blood products.
So no changes in trauma care for now, and sadly no astoundingly impressive evidence for TEG/ROTEM changing outcomes from haemorrhage in trauma BUT lots of food for thought and potential directions for new research, especially in those prespecified sub-groups (coagulopathic and head injury) – and while point-of-care testing might not be the tactic we were looking for, our strategic care of trauma patients seems to be heading in the right direction – and that’s a win, after all.
- Viscoelastic haemostatic assay augmented protocols for major trauma haemorrhage (ITACTIC): a randomized, controlled trial https://link.springer.com/article/10.1007/s00134-020-06266-1
- JC. ROTEM for ED coagulopathy detection. St.Emlyn’s https://www.stemlynsblog.org/jc-rotem-ed-coagulopathy-detection-st-emlyns/
- Why are bleeding trauma patients still dying? https://link.springer.com/article/10.1007/s00134-019-05560-x
- BLEEDING AND COAGULATION RESEARCH PROGRAMME https://www.c4ts.qmul.ac.uk/research-programmes/bleeding-and-coagulation
- Thromboelastogram (TEG) https://litfl.com/thromboelastogram-teg/
- JC: Why do bleeding trauma patients die? St Emlyn’s https://www.stemlynsblog.org/jc-why-do-bleeding-trauma-patients-die-st-emlyns/
- The use of viscoelastic haemostatic assays in the management of major bleeding https://onlinelibrary.wiley.com/doi/full/10.1111/bjh.15524