Administration of tranexamic acid (TXA) is integral part of the management for major trauma patients in the UK. It is deemed such an important aspect that its administration is used as a quality indicator to benchmark major trauma centres across the country. TXA is a medication that has raised a fair degree of controversy, mostly in the US, and whilst the current evidence base is not perfect (Ed – there is no such thing as a perfect evidence base outside of n=1 trials), we believe that until new evidence arives to the contrary then we should be giving it to our major trauma patients.
If you go back to the original CRASH 2 paper you will know that they enrolled patients up to 8 hours post injury. That’s a pragmatic approach, but with many bleeding deaths from trauma taking place before that it does not quite fit with the pathophysiological model. In that study there was a strong signal that earlier treatment benefited patients more, but how can we look at this from a wider perspective? Arguably a secondary analysis looking at dividing patients into time bands from more than one trial, as to when they get the TXA might suggest whether there is a time to treatment effect.
This week a really interesting article in the Lancet1 (you can access it here) reinforces the pivotal role and safety profile of this drug in the prevention of traumatic bleeding, and attempts to further investigate the time to treatment effect in TXA.
The St.Emlyn’s team have had a look and summarised this promising paper for you but as always we do suggest you read the original article in full before you draw your own conclusion(s).
What we know on the topic:
- Acute severe bleeding is a leading cause of death in the developed world
- Antifibrinolytic drugs reduce bleeding by inhibiting breakdown of fibrin clots
- Antifibrinolytics reduce surgical bleed and the need for transfusion
- Administration of tranexamic acid (TxA) within 3h of bleeding onset reduces deaths from bleeding in trauma 2 and postpartum haemorrhage3
What we did not know so far:
- The effect of treatment delay on the effectiveness of antifibrinolytics in patients with acute severe bleeding was controversial (to some), or rather we could not be certain from the studies that are already in press.
What did the authors look at?
- They examined the effect of treatment delay on the effectiveness of antifibrinolytics
- They did so by conducting an individual patient-level data meta-analysis of randomised placebo-controlled trials done with more than 1,000 patients. Individual patient meta-analysis is a pretty coool approach where instead of combining the pooled results of trials, the authors are able to dig into individual patient data and pool them as one study. This is arguably more powerful and more reliable than simply pooling trial summaries4.
How did they do it?
- They identified trials undertaken between 1946 and 2017 using a register of seven databases (Embase, MEDLINE, CENTRAL, Web of Science, PubMed, Popline, WHO Clinical Trials Registry)
- Two authors screened the abstracts and discrepancies were solved by consensus
What were the considered outcomes?
- Primary outcome: death from bleeding. This is interesting and a bit controversial. Deaths from bleeding are a subgroup of all deaths and it’s possible that there would be a bias towards a positive TXA effect by doing so. Whilst we understand why they went for bleeding deaths as a primary outcome we think it’s very important to report overall mortality too. They did do this, but it’s a bit tricky to find.
- Secondary outcomes: vascular occlusive events fatal and non-fatal (myocardial infarction, stroke, VTE)
- Treatment delay was estimated as the interval between the bleeding onset and start of fibrinolytic treatment. It must be noted that the CRASH-2 trial reported treatment delays but the WOMAN trial estimated the delay as the interval between birth and randomisation
What data were presented?
- All analyses were done according to an intention-to-treat principle
- Treatments effects were expressed with odds ratios (OR) and 95% confidence intervals (95% CI)
- Homogeneity of treatment effects was analysed (please see the original article and appendix for full details on this)
What were the results then?
- 12,573 records were screened, with only two included in the final analysis after exclusions
- The CRASH trial assessed the effects of TxA in 20,211 bleeding trauma patients and the WOMAN trial the effect of TxA on death, hysterectomy and other morbidities in 20,060 women with post-partum haemorrhage
- Individual patient-level data were therefore obtained for 40,138 participants from the above-mentioned trials: 20,094 received TxA and 20,044 placebo
- Of the 3558 deaths, 1,408 (40%) were due to bleeding, of which 884 (63%) occurred within 12h of bleeding onset
- In the WOMAN trial, deaths from bleeding peaked at 2-3h after bleeding onset
- TxA significantly increased overall survival from bleeding OR 1.20, 95% CI 1.08-1.33, p=0.001.
- You have to dig into table 2 to find out what that means in terms of real patient difference and it’s an absolute risk reduction difference of about 0.6% (Ed – a much better way of describing it). That’s interesting and perhaps a little less dramatic as a result . As you know, an absolute risk reduction of 0.6% means that the number needed to treat is about 166 for all bleeding mortality.
- It was estimated that the treatment benefit benefit decreased by 10% for every 15 minutes of treatment delay, and so if you are looking for a significant clinical effect, early administration seems to be the way forward.
- The risk of occlusive events was higher in trauma patients when compared to the post-partum haemorrhage group
- There was however no increase in fatal vascular occlusive events with TxA (OR 0.73, 95% CI 0.49-1.09, p=0.1204) with no heterogeneity observed between trials
What were the drawbacks?
- Studies with fewer than 1,000 patients were excluded as it was felt they are underpowered to assess effects on death and there is a risk of selective reporting
- The time of death was available only for the post-partum haemorrhage group. The distribution of death was however similar between the two groups
- Deaths due to bleeding and those due to severe vascular events (like DIC) could have been misclassified as they have similar pathophysiologies
- There is an argument whether the physiology of bleeding between the two groups varies
What were the authors’ conclusions?
The authors concluded that death from bleeding occurs soon after onset and even a short delay in treatment reduces the benefit of TxA administration.
This meta-analysis is not without some flaws as detailed above but it appears to further establish the role and safety profile of TxA in the management of major trauma patients.
TxA is cheap, readily available and easy to administer so every trauma patient at risk of bleed should get it, right?! 🙂
Or, is the pragmatic question not what will you folks change in your daily practice in order to ensure that TxA gets to your trauma patient sooner rather than later?
@baombejp on Twitter