Tranexamic acid in trauma is a well-established standard of care in most mature trauma systems. The two major trials of TXA, CRASH 2 and 3 suggest a small but important decrease in mortality for patients with significant bleeding and in subgroups of patients with head injury. We’ve looked at TXA in some depth on the blog, and also dabbled in some of the controversies around whether it works (we also remain skeptical in some areas). That said it is a widely used drug in the UK.
This week we look at a paper in the BJA that examines whether there are gender difference in the administration of TXA to patients who may potentially benefit. This is an important topic as there is evidence from many other trials that clinicians are biased in the treatment of patients as a result if gender or ethnic differences. This is obviously something that we should try and understand and thus avoid. The paper abstract is below, but as always we recommend you read the full paper. For this paper it’s especially important as it’s really quite complex.
What sort of paper is this?
It might be easiest to think of this as two papers in one.
Paper one looks at the CRASH 2 and 3 studies to see if gender affects the effectiveness of TXA in major trauma.
Paper two looks at the TARN databased to see if there is a gender difference in the administration of TXA based on gender.
Tell me about the gender difference in TXA effectiveness
In this first part of the paper, the authors went back and looked at the data from CRASH 2 and 3. They performed a secondary analysis on a per-patient basis. This is relatively straightforward, although we must remember that it is a secondary analysis that should be considered hypothesis generating.
The data suggests that there is unlikely to be a sex related difference in the effectiveness of TXA. That seems biologically plausible to me and seems reasonable to accept although there are significantly wider confidence intervals for the female cohorts as opposed to the males ones (smaller numbers). One strength highlighted by the authors is the large number of patients in the combined trials. Arguably this would be further strengthened if it included data from other trials of TXA (and perhaps we will see this soon).
It’s worth remembering what the inclusion criteria for CRASH 2 and 3 were at this point as this is very relevant later.
- Inclusion: adults (age > 18) with trauma; present within 8 hours of incident; either significant haemorrhage, or who are considered to be at risk of significant haemorrhage (systolic blood pressure < 90 mmHg and / or heart rate > 110 bpm); responsible doctor was substantially uncertain about whether or not to treat with TXA
- Exclusion: clear contra-indication to TXA
- Adults with TBI within 3 hrs of injury (changed from within 8 hrs of injury in Sep 2016)
- GCS <=12 or intracranial bleeding on CT
- No major external bleeding
- The treating clinician had to be substantially uncertain as to the appropriateness of TXA as a treatment
I’ll leave the first part of the paper there as it’s not the most interesting or challenging part of the paper, but suffice to say that there is no clear evidence of any difference in the effectiveness of TXA between the sexes in these trials. On that basis it seems reasonable to conclude that we should be giving TXA to the same sort of patients irrespective of whether they are male or female.
Tell me about the administration question.
The second part of this paper looks at whether there are prescribing differences for TXA between gender.
Tell me about the dataset.
The authors have gone to the TARN database as a dataset as this records data on trauma patients from submitting hospitals and importantly records whether the patients received TXA.
Eligibility for TARN is a little complicated but essentially it aims to capture trauma patients who require significant admission to hospital with specific injury patterns. More detail on TARN eligibility can be found here.
What’s the outcome?
This is a really important point in this study as if we are to determine whether there is differential prescribing of TXA then we need to understand what we mean by appropriate prescribing. In this study TXA is deemed to have been indicated for any patient who is in the TARN database with an ISS>9. This is (I think) based on the presumption that all these patients will benefit from TXA. Whilst there is an association with TXA use and reduced mortality with increasing ISS, we must remember that this was not the inclusion criteria for CRASH2/3 on which the evidential base is derived for equity of benefit in the first part of the study.
For clinicians the decision to prescribe TXA is based on the information available to them at the time they prescribe it. ISS is not available to clinicians at that point in time.
ISS in itself is complex here too as there are many ways to get to an ISS of >9 which have significantly differential rates/risks of severe bleeding
What did they do?
In very basic terms the authors looked at the TARN database and identified patients with an ISS>9. They then looked at whether or not the patients received TXA and whether this differed depending on gender.
The analysis was adjusted for a number of factors, notably age and the BATTS which is a score of bleeding potential. BATTS was developed from the TARN database and has some crossover in authorship with this paper.
Interestingly the authors did not adjust for mechanism of injury and the BATT score has a limited adjustment for this. This does mean that there appears to be significant discrepancies in the data sets with regard to the type of trauma encountered. Female patients were older and less likely to have high energy trauma mechanisms. As a result the overall risk amongst females was lower than that of males. The authors suggest that including MOI would over-adjust but I do think it may have an influence on clinical decision making (i.e. the point at which clinicians decide to prescribe or not).
The headline figure is that TXA was received by 7.3% of females and 16.8% of males. That’s quite a difference, and likely to be clinically important.
The critical appraisal question is whether the comparison groups are comparable. There does appear to be significant differences between the two groups that extend beyond gender and these may bias the results. Low energy falls in the elderly probably have a differential TXA benefit profile. Similarly the decision to use TXA in the initial resuscitation phase is influenced by the mechanism of injury. It may (it is) harder to spot significant injury in low energy falls as compared to high injury mechanisms, and since a major cognitive trigger to using TXA is the mechanism this may account for some of the difference.
The authors have adjusted the results using the BATT score in an effort to adjust for potential confounders. This shows that the differential administration exists at all levels of BATT risk, but that the differentials are less significant with higher mechanisms of injury. If you look in the supplementary data there is evidence that the sex differences increases with advancing age.
It appears that the group with the biggest differentials are the elderly, low energy falls, with minimal physiological change. The least differentials are in the younger, high energy, physiological abnormal patients. This makes intuitive sense to me based on how I see TXA decisions being made in practice. Note that it does not justify them, it’s just that this meets my preceding thoughts and biases, and even if the differential at lower ages is less, this is still an important finding.
This secondary analysis confirms that TXA is probably equally effective in male and female patients. It also suggests that TXA may be differentially prescribed between males and females, that certain groups are more at risk of this differential. However, it cannot tell us why this is, what the impact of that differential is, nor help us determine what to do next. These are areas I’d love to hear more about in future. I’d also like to see a similar analysis done around other protected characteristics relevant to our practice.
In short, we don’t know whether the difference found is due to the fact that the patient characteristics are different, or whether there are systematic behaviours that lead to differential prescribing. However, we can see that there is a difference and that is worthy of investigation. Whichever it is, I’m hopeful that we can understand more on this issue in the future.
Use of tranexamic acid in major trauma: a sex-disaggregated analysis of the Clinical Randomisation of an Antifibrinolytic in Significant Haemorrhage (CRASH-2 and CRASH-3) trials and UK trauma registry (Trauma and Audit Research Network) data https://www.bjanaesthesia.org/action/showPdf?pii=S0007-0912%2822%2900184-2
BATTS score https://bmjopen.bmj.com/content/9/5/e026823
Ashley Liebig, “Gender, conferences and careers in EM. St.Emlyn’s,” in St.Emlyn’s, August 29, 2016, https://www.stemlynsblog.org/gender-conferences-careers-em-st-emlyns/.