Over the years we have had more than a passing interest in tranexamic acid. In part because we have been involved in some of the research, recruiting to trials, or acting as principal investigators, but also because it’s a drug that so often sits on the cusp of analysis and where the principles of evidence based medicine are easy to see and debate. Having looked at TXA in trauma, childbirth and GI bleeding it’s now time to look at it’s use in epistaxis with the recently published NoPAC study.
TXA for epistaxis has been widely advocated, but the evidence for this has never been particularly strong and we have remained sceptical as trials such as HALT-IT showed that TXA is not a panacea for all forms of bleeding and we cannot simply extrapolate the results of trauma trials to nose bleeds. It’s just a different sort of bleeding and therefore we need different trials.
Fortunately just such a trial (the NoPAC trial) has been completed and is now published in the Annals of Emergency Medicine. You can read the abstract below, but as we always say, please read the full paper and make your own mind up.
What sort of study is this?
This is a randomised controlled trial which is the best design to test an intervention such as TXA or nasal packing.
Who was studied?
The trial was conducted across multiple sites in the UK between 2017-2019 with Virchester being one of the sites. Adults with persistent epistaxis in the ED were eligible for inclusion in the study, but only if simple first aid measures had failed to stop the epistaxis.
This is important as the simple aid measures (squeezing the nose using a nose clip or the use of ice) actually stopped a lot of bleeding and therefore rendered the patients ineligible. The simple aid measures were applied for 10 mins and only if persistent bleeding existed after this then randomisation took place.
What was done?
If simple measures failed then a topical vasoconstrictor was applied into the nose. We did this using a dental roll soaked in the vasoconstrictor and then applied that to the bleeding nostril and held it in place with a nasal clip. Only if the patient was still bleeding after the vasoconstrictor and clip was the patient randomised into the trial of TXA.
It’s important to realise that the TXA randomisation was done after a significant amount of routine intervention had already taken place and in fact a large number (143 patients) stopped bleeding after vasoconstrictor therapy. The choice of vasoconstrictor was a local decision, but we understand that many sites (including us) used phenylephrine, others used adrenaline and perhaps others used something else.
If they were still bleeding after the vasoconstrictor application then in the NoPAC trial the dental roll was exchanged for one soaked with either 2ml of TXA or with a placebo. This was again left in place for 10 minutes. If bleeding persisted then another 2ml of TXA or placebo was applied, again with a nasal clip.
If this failed then the patients had formal nasal packing to control bleeding.
What about the outcomes?
The primary outcome was very simple, did the patient stop bleeding. If they did not stop then the patients had nasal packs inserted and this was deemed a treatment failure. A number of secondary outcomes were examined such as recurrence, admission, need for transfusion etc.
What are the main results?
The NoPAC study screened 2622 patients, but due to a variety of reasons (including just stopping bleeding as a result of simple measures or vasoconstriction), they ended up with 496 patients randomised to TXA or placebo.
The headline figure is that 111/254 (43.7%) of patients treated with TXA needed to be packed, vs. 100/242 (41.3%) of patients treated with placebo. This was not statistically significant and not clinically significant (odds ration 1.11 95% Confidence Intervals 0.77-1.59). Similarly there was no difference in secondary outcomes, although there were a number of serious events, but with small numbers of patients affected. However, there was no real signal that the adverse events were trial related.
There is a nice summary from @Lwestafer below
What does this mean?
On the current evidence the use of TXA for epistaxis cannot be recommended as it seems very unlikely that there truly is a difference between the therapies. However, there are a few words of caution.
When the study was presented at the RCEM conference I remember the authors describing how they were surprised at the low packing/failure rate in the placebo arm. The original power calculation had estimated a much higher rate and so the study may be underpowered, but perhaps this also shows that simple measures work more effectively than we had anticipated. Despite the potential for it being underpowered there really is no signal of benefit though.
Patients were not recruited sequentially in the study and so it is possible that this is a patient group affected by unknown biases. The study did not differentiate between anterior and posterior bleeding as it was argued that it was too tricky to do this in the ED. I have some sympathy for this perspective, but as posterior bleeds were unlikely to be successfully treated with anterior treatments, a small difference in posterior vs. anterior bleeds between the groups may have affected the results (though pragmatically I think the authors approach is the correct one). The authors suggest that more research might be of benefit using different doses, or different application methods.
Since publication there has been a vigorous debate on twitter about the results. As with any TXA study there has been some disappointing comments about TXA as a drug in general across all conditions which makes little sense. It’s important to test TXA in different studies for different indications and not try to suggest that this trial tells us anything about whether it is effective in traumatic haemmorhage or any other form of bleeding. Our view is that we take each trial on it’s own merits. TXA shows no sign of benefit in epistaxis, nor in GI Bleeding, but does in bleeding trauma, and in others such as head injury the picture is not as clear (but pragmatically we recommend it in some subgroups). The point is that bleeding is complex and we should take each study on its merits. For some reason (which I still cannot quite fathom) TXA is a polarising drug that seems to create quite powerful anti-TXA sentiments with people expressing hate, or sometimes love for TXA. This is utterly irrational. We should not have an emotional attachment to a drug, but rather we should look at the evidence and make rational/pragmatic decisions. Personally, and in keeping with many of our great friends (for example Ken Milne) I remain sceptical about TXA in all conditions but will act on the best available evidence until something better comes along. Interestingly, of the last two TXA trials I have recruited to, both have been negative (and for HALT-IT I was a principal investigator). In those areas where the evidence is less strong it is reasonable for clinicians to come to different conclusions about what they would do with patients (on the balance of probabilities), but that has always been the case in medicine, and the same should be said of TXA.
Mark Ramzy has a nice infographic on this that he has helpfully updated as below (read the second one which is the update).
I was also interested to read that many clinicians appear to have adopted TXA as a standard of care prior to NoPAC results. Personally I never adopted TXA as a treatment for epistaxis as I wanted to wait for the evidence and to encourage the enrolment of patients in NoPAC. However, others did on the basis of lower quality evidence and must now reconsider whether they wish to continue (I think they should stop). There is no shame in this as our ability to be agile in our thinking and practice is a cornerstone of EBM.
There have also been many comments about the admission rate for epistaxis in this study which at 45% felt very high to many readers.
Our US friends were concerned at the number of patients admitted with epistaxis. This is a fair point and worthy of analysis. There are several reasons for this.
- The UK is a different health economy and it is likely that many patients with epistaxis will not attend EDs such as the ones recruiting in the NoPAC study as alternatives are available. So we may have seen a more severe cohort at presentation.
- Many patients were excluded from recruitment as simple measures worked in the majority of patients. Thus this is not a general group of patients with epistaxis, but rather those that we cannot control with usual methods (so a more difficult group to deal with).
- Custom and practice in many UK centres had been to admit any patient requiring a nasal pack to the ENT service. Our US colleagues state that this is not the case in their practice and that’s really interesting and useful information. It is clearly time for UK clinicians to look at nasal packing as an ambulatory intervention (and anecdotally similar examples have been happening as a result of the COVID-19 pandemic). The latest COVID-19 guidelines from the ENTUK explicitly now state this as a response to the pandemic.
The bottom line.
Simple measures, combined with topical vasoconstrictors work well at controlling epistaxis in patients presenting to the ED.
Reuben A. et al The Use of Tranexamic Acid to Reduce the Need for Nasal Packing in Epistaxis (NoPAC): Randomized Controlled Trial. https://pubmed.ncbi.nlm.nih.gov/33612282/
Chris Gray, “JC: Halt! It’s not time for TXA! Or is it? HALT-IT results at St Emlyn’s,” in St.Emlyn’s, June 23, 2020, https://www.stemlynsblog.org/halt-it-st-emlyns/.
COVID 19 ENTUK guidelines for the management of epistaxis https://www.entuk.org/sites/default/files/COVID%2019%20Epistaxis%20Management.pdf
Simon Carley, “JC: Tranexamic Acid (TXA) in Head Injury. The CRASH-3 results. St Emlyn’s,” in St.Emlyn’s, October 14, 2019, https://www.stemlynsblog.org/jc-tranexamic-acid-txa-in-head-injury-the-crash-3-results-st-emlyns/.