The TTM trial made the assumption that it was the avoidance of fever rather than hypothermia per. se. that was beneficial, and used this assumption in a slightly odd way….
They compared 33 to 36 degrees as temperature targets, which is an interesting choice. Why 33? That ‘dose’ is the same as the two previous human studies, but was derived from three prior studies; one in gerbils, one in dogs and one in rats – fair enough. 36 degrees is a really interesting target. It is not physiological, and shows that this is NOT a trial of cooling versus not cooling but of two different doses of cooling, one of which is entirely novel. This is an odd leap and confuses the question somewhat in that there is not really a control group, just 2 intervention groups; one that is already in question (33) and another that is completely new (36). This is also the first trial of different hypothermia targets in humans.
There’s lots as always to talk about with this paper, but just to bring out a couple of other points….
I always find mortality an interesting concept in trials such as this, as unless brain death occurs we are really looking at withdrawal of active management, which is multifactorial and complex. In designing this study, the authors needed to make some attempt to standardise this decision making whilst still maintaining the autonomy of the responsible clinician (and avoid criticisms of a lack of blinding). What they came up with is elegant but also sounds pretty brutal! (explained in detail in the supplementary appendix). They defined a list of criteria which would result in a recommendation to withdraw treatment, assessed by an independent (and blinded) neurologist. There was no requirement to follow this recommendation. By the back door, the authors have given us a set of criteria we could potentially use in our practice. They are; brain death, within the first 24 hours the presence of myoclonic status with absence of the N20 SSEP*, or at 72 hours extension to pain with either uncontrolled status epilepticus or absence of the N20 SSEP. I’ve recently become worried about using myoclonus as a clinical sign after reading this however. We don’t routinely do SSEP, maybe we should….?
The study had power to detect a 20% RRR, which with mortality running at approx. 50% would equate to a 10% absolute risk reduction i.e. in a group of 100, 60 people would survive rather than 50. The question is therefore; would you still institute a therapy if it only saved 5 lives rather than 10? Where would be your cut off? What would you take into account when deciding? A smaller difference of course requires a bigger sample…
The punchline is that there was no difference, either in mortality, adverse events or neurological outcome. Should we abandon therapeutic hypothermia? If not to what temperature should we cool? Is it that 36 degrees is an adequate dose of therapeutic hypothermia and 33 gives no added benefit? Does this study show that we shouldn’t limit to VF and VT? Is fever control the answer? – I know what I think about these questions and how this study informs the argument, but I’d be keen to hear what you think
– Please leave a comment below………….