Friday, 30 June 2017

Prehospital administration of TXA in trauma patients: a hypothesis raising study

The CRASH-2 trial put tranexamic acid on the map for patients with trauma at risk for significant haemorrhage. Although the reported 1.5% absolute mortality reduction was impressive, there are some concern as most of the data was obtained in resource poor countries. Is this externally valid to us? Perhaps not.

So, more data is needed.

This German study was a retrospective look at a database and registry to try to determine if TXA could be helpful when given prehospital. It sought to compare sick trauma patients who happened to get TXA vs. those that did not. They looked at a whole lot of outcome measures.

Its design included some fancy matching of cohorts and some complex propensity scoring.

Without reading any further, it is obvious this design could never claim cause & effect. At best it would be hypothesis generating. Too many assumptions, confounding and making decisions based on poor quality registry data.

For what it’s worth, they found a big early mortality difference with TXA. It was 5.8% vs. 12.4%. Let’s face it, this is too good to be true. More believable was overall hospital mortality that was similar in both groups with an absolute trend of 1.5% benefit in the TXA group… Wait a minute, where have we seen this before? CRASH-2 anyone?

The authors conclude, “until further evidence emerges, the results of this study support the use of TXA during prehospital treatment of severely injured patients.”

This conclusion is a stretch if only looking at this paper. But given other data on TXA, I think the overall message is reasonable.

Good news; we will probably be getting much better data soon. The PATCH trial is estimated to wrap up in December of this year. This large multi-centre double blind RCT of TXA is what we genuinely need to answer the question.

So why did we need to raise the hypothesis?


Wafaisade A, Lefering R, Bouillon B, et al. Prehospital administration of tranexamic acid in trauma patients. Crit Care. 2016;20:143.

A meta-analysis of amiodarone or lidocaine for cardiac arrest: when crunching numbers is not appropriate

No anti-arrhythmic drugs have been proven to result in increased meaningful outcomes after cardiac arrest. Sure, a larger proportion of patients arrive at the hospital alive but they all die anyway.

A recent RCT of 3026 patients published in 2016 reported about a 3% trend towards improved survival to hospital discharge with amiodarone or lidocaine as compared to placebo. If this trend was true, it would be rather important. A NNT of 33 for survival is nothing to shake a stick at!

Perhaps a systematic review & meta-analysis with more statistical power give us the answer? Why not.

These well intended authors conducted such a study.

As is often the case, their search found tons of papers but in the end only 3 RCT’s.

One RCT had 3026 patients… sound familiar?

The other two only enrolled 347 and 304 patients respectively. To be clear, 80% of the data comes from one study.

So, the authors quit here and acknowledged they shouldn’t go much further until such time more evidence became available.

Just joking…

Of course they crunched some numbers and came to some conclusions. As expected, it didn’t really provide us with any new information.

To be fair, their overall methods were good as was the discussion. I thought their advocacy for the reconsideration of lidocaine was interesting and probably correct. It’s great when things come full circle.

In the end, the jury is still out. But we knew this already… with no need for crunching.