Apologies if the stats here seem either too arcane or too obvious- my
defence to being too obvious is that I didn't know them as a registrar; my
defence to being too arcane is that I think they are at the root of how
you'd design the study being proposed.
> The first question that I feel needs answering (accepting
> that survival is
> clearly the most important outcome measure) is that
> acceptable levels of
> oxygenation are achievable when compared to conventional
> single rescuer BVM
> CPR. Standardising this will clearly be a challenge.
That's going to be harder to get through a REC- if you come at it from the
angle that you're trying to see if this treatment is as good as standard
treatment, it is a life or death situation and you don't have patient
consent (basically you think you've got a worse treatment but you don't
think it's significantly worse). Other problem with this is the stats-
pretty difficult to prove that a treatment is as good as another one unless
it's better. You have to set a higher beta value than 0.8 because beta error
is more important. I suppose it gets down to what is a clinically important
difference that you want to exclude. So if you're looking at a clinically
significant difference being 0.25 SDs (which covers about 10%) you need
around 250 patients for an 80% chance of picking up a difference at p= 0.05;
but too exclude something you probably want a beta of at least 0.95%
(needing a bit over 800 patients), and probably 0.99 (needing about 1,200).
Despite the fact that cost effectiveness is considered important these days,
you'll be lucky to get a REC to consider a study of a treatment presumed to
be inferior on the basis that it's cheaper (looks like you're comparing this
method of CPR to putting another paramedic on the wagon)- especially without
informed consent. I think it sounds better pitched as a way of avoiding the
positive intrathoracic pressure and thus improving cerebral perfusion with
any effect on oxygenation unlikely (given existing evidence)
> It occurs that some of this data might be usefully obtained from
> irretrievable resuscitation scenarios with blood gas analysis with
> conventional compared with AED continuous chest compressions.
That's going to be a really tricky one to get past a REC in the wake of the
Alder Hey business- getting RECs to approve experimenting on the recently
deceased without discussion with relatives is difficult. Most relatives
would probably be OK with it, but if even one was to kick up a fuss
(possibly looking for a pay- off). However, it is possible to get
permission for non therapeutic research without consent provided it doesn't
harm the individual in certain circumstances. I think the issue here is more
legal (doing something to someone that will not benefit them and without
their consent) than ethical. It does seem a bit analagous to ventilating
potential organ donors who have no chance of survival themselves (but
arguably worse, because the benefit to others is less clear cut)- legal in
the US, not in the UK. You might just swing it if you waited until the
relatives got in and them asked them, but still tricky.
> First of all when you mention ventilation are you referring
> to active ventilation such as mouth to mouth, BVM in a basic
> CPR delivery expecting that chest compression itself will
> cause enough change in thoracic pressure to generate some
> flow of air and therefore oxygen delivery?
Yes. This has been shown to occur in experiments.
> I am unsure if your comment is meant to include ET intubation
> in these situations or not.
Anything that keeps the airway open will do.
> It occurs that some of this data might be usefully obtained from
> irretrievable resuscitation scenarios with blood gas analysis with
> conventional compared with AED continuous chest compressions.
>
Can someone clarify a bit about the technicalities of blood gas analysis
during CPR: How do you know you're in an artery? (I'm never sure whether
what I feel during CPR is an arterial pulse of transmitted pulse wave down a
vein)? Does it take long to get in?
Given that the papers cited earlier show that blood oxygenation doesn't
change much whatever you do about ventilating, isn't cardiac output more
important?
> Also, wouldn't positive thoracic pressure be transient and
> decreased capillary blood flow more of a theoretic rather
> than actual concern? I can't imagine it making a great deal
> of difference in survival outcomes.
Even without a PEEP valve, passive ventilation with ETT (and indeed BVM)
always has some intrathoracic pressure (obvious during inspiration; but
expiration won't occur unless there intrathoracic pressure is above
atmospheric. This will reduce blood flow into the thorax, with back pressure
reducing perfusion of extrathoracic organs. With the already low perfusion
pressures achieved from CPR, this could prove critical in terms of brain
perfusion (some patients will have adequate CPP even with positive
intrathoracic pressure; some inadequate even without; but there are probably
a few in the middle in whom those extra couple of centimetres of water make
the difference to whether any blood flows through the capillaries.
This is certainly a (pathophysiology) question that remains open to debate:
given that positive pressure ventilation is likely to decrease perfusion
pressures in vital organs, at what point does the detriment outweigh the
benefit? Would there be any benefit from oscillatory ventilation or similar?
The clinical question is: given that a large number of studies have shown no
difference in outcome, does it ever matter? And if so, can we identify which
patients will benefit from which treatment?
Matt Dunn
Warwick
This email has been scanned for viruses by NAI AVD however we are unable to
accept responsibility for any damage caused by the contents.
The opinions expressed in this email represent the views of the sender, not
South Warwickshire General Hospitals NHS Trust unless explicitly stated.
If you have received this email in error, please notify the sender.
|