A new study in the BMJ this week suggests benefit from use of the new oral anticoagulants. This is of significance to those of us in UK emergency medicine involved in the diagnosis and initial treatment of thromboembolic disease. Our current practice is for patients with a diagnosis of PE or DVT to be started on low molecular weight heparin (LMWH) until the patient is established on warfarin. This is at first glance a fairly straightforward process, but politics, patients and resources still get in the way. It is not always that straightforward to get patients into the anticoagulation clinic if they are from outside the local area, if they have intercurrent illness or if the service is resource challenged. Under a myriad of circumstances the ED ends up picking up the pieces and so it is something we cannot avoid.
So, the new oral anticoagulants Rivaroxaban, Dabigatran, Apixaban and Ximelagatran are out there and in the thromboembolic world there is increasing interest in using these drugs for the treatment of acute VTE events. Well Ximelagatran won’t as it has been withdrawn from the market, but there may be others on the way.
So the BMJ article is worth a look as these drugs are coming our way fast. The paper edition has the brief report, online is the full meta-analysis and systematic review.
The headline numbers here are firstly that the trials cover over 16000 patients (a lot), and that there is no significant difference in the rate of recurrent DVT between the new oral anticoagulants and traditional anti Vit-K therapy (warfarin & sinthrome). That’s great in terms of patient convenience but emergency physicians like us are not just interested about when things go well, but also when they don’t and there has been a lot of concern about the management of bleeding secondary to the new anticoagulants. In the BMJ article the pooled data suggests that the new oral anticoagulants have a lower complication rate as compared to traditional anticoagulants. In the BMJ paper edition they report relative risks (which are very unhelpful!).
If we look at how the data is reported in the paper edition we see results that for major bleeding…
- Rivaroxaban has a relative risk of 0.57 (CI 0.39-0.84)
- Dabigatran has a relative risk of 0.76 (CI 0.49-1.18)
Wow! At first glance that looks great. The reduction in risk of major bleeding is a huge amount… or is it? In reality these figures are almost meaningless unless we also take the event rate into account. Let’s face it, if the event rate is ‘almost never’ then the reduction to o.57 of almost never is 0.57 the rate of ‘almost never’ – in other words it is of no clinical importance. So, to get any meaningful data we need to go to the online version of the paper and look at the event rates of major bleeding in each group, from there we can work out the number needed to treat to avoid major bleeding. So, if we go digging into the data online to look at major bleeding rates we find…
- Rivaroxaban has an event rate of 43/4382 vs 74/4379 for trad methods.
- Dabigatran has an event rate of 35/2553 vs 46/2554 for trad methods
So, we can then see the event rates which are small. Around the 1-2% mark in fact. So any reduction is going to have a relatively small impact on a patient population. Again, if we have the time to do this ourselves we can work out the number needed to treat by comparing the combined event rates for these two drugs.
- The combined event rate for the new oral anticoagulants (major bleeding) is 1.1%
- The combined event rate for the old oral anticoagulants (major bleeding) is 1.7%
Not so impressive now is it?!
Even better – we can convert these event rates into numbers needed to treat and it comes out as 0.6 (the absolute rate reduction) divided by 100… which is a whopping NNT of 166 patients! A figure which is high and something that ‘feels’ very different from the reported risk reductions in the paper.
We can also play the same game with the overall mortality figures. If we do that we find that in the new oral anticoagulant group the combined rates are…
- The combined event rate for the new oral anticoagulants (all cause mortality) is 2.1%
- The combined event rate for the old oral anticoagulants (all cause mortality) is 2.1%
I don’t think you need to do the maths on this one. The NNT is basically infinite.
At this point I am getting concerned. The paper rightly concludes that there is no difference in all cause mortality but also points to the difference in major bleeding reduction. However, as clinicians we need to ask ourselves about the clinical significance of the difference in major bleeding rates, and clinically why am I bothered? Probably because rather selfishly when it goes wrong these patients are likely to end up in my emergency department and someone has to reverse the anticoagulation. That someone is likely to be me and whilst I know how to reverse warfarin, reversing the new anticoagulants is a whole different proposition.
I don’t quite know why but I am increasingly concerned about our direction of travel with the new agents. It seems inevitable that we will (and are) see them in practice, but the evidence base for their safety and efficacy in practice as clearly outlined in this paper (the online version of the paper has a very good critique of the limitations of the baseline studies) are more concerning than I had previously realised.
I also ran this blog post past the @thegreathornero – the resident expert on all things VTE (on leave of absence from St.Emlyn’s at the moment) – and he made three really valid points.
1. Firstly, there are concerns about how we are comparing the NOACs against warfarin as a measure that we need to address is the time in theraputic range (i.e. how often are you patients really anticoagulated for). Surprisingly it is often quite low (60% quoted recently) which questions whether warfarin is a valid control, or should it be typically controlled warfarin or well controlled warfarin!
2. Secondly, we need to understand what is meant by major bleeding as it’s one of those terms that we sort of know what it means to us, but do we all agree? Well fortunately there is an agreed definition which can be found in the Journal of Thrombosis and Haemostasis, and which is shown below. It’s clear that major bleeding is pretty serious stuff by these criteria, but it’s good to know that we have agreed definitions (if you are interested there are also agreed definitions for cardiovascular trials here which, interestingly, are different). I think we would also agree that these are the sort of situations where reversal would be really important so again I start to get nervous about those reversal strategies again.
- Fatal bleeding, and/or
- Symptomatic bleeding in a critical area or organ, such as intracranial, intraspinal, intraocular, retroperitoneal, intra-articular or pericardial, or intramuscular with compartment syndrome, and/or
- Bleeding causing a fall in hemoglobin level of 20 g L−1 (1.24 mmol L−1) or more, or leading to transfusion of two or more units of whole blood or red cells.
3. Lastly, that there is actually a much better paper than anything I can put together on pretty much the same subject, and with a similar conclusion in the Annals of Internal Medicine this month.
As Emergency Physicians we will need to know about these agents for two reasons. Firstly, we will need to work closely with our haematology colleagues to ensure that we start patients with confirmed VTE on the right agents. Secondly, the increased use of these agents means that we need to become proficient at the management – and in particular the reversal – of these agents when major bleeding does occur. The latter is a whole other story that I am yet to get my head around. We need a blog post on it asap…. (but it looks complex, so volunteers welcome).
Finally can I plug the EMJ (Emergency Medicine Journal) podcast where Dan Horner interviews Phil Wells (yes, that Wells of the scoring fame!) on VTE, scoring systems and the future of VTE research. Listen and learn from current and future legends of VTE research.
Simon Carley & Dan Horner