Journal club was interesting this week. The marvellous AR brought a paper along from the EMJ looking at the use of non-invasive tissue oxygen monitoring in septic patients. Aha, interesting we thought, there is certainly something that appeals to us about the use of a monitor that might actually tell us about tissue perfusion. This is really on the back of a long running conversation we have in the resus room about the difference between blood pressure and flow. The conversation often ends with me threatening to inject the other person with 4mg of Noradrenaline to see if they still think that blood pressure is the most important measure of tissue perfusion – nobody has taken me up on this, but you get the idea.
Anyway, the point of the argument is that many of the basic non-invasive measures of cardiovascular performance (such as BP) are really a bit rubbish. We know this, we teach this, we do other stuff to work out what is happening (more of this later), but there is perhaps some appeal for us to get an idea of what is going on at tissue level in terms of perfusion.
So, there are machines around that purport to measure tissue oxygen saturation such as the Inspectra StO2 as used in the study we were looking at. Now, I have not used this machine and have no links to the company in any way, I’m just saying that such machines exist and that on face value there may be some appeal to the concept…, now, to the paper. Vorwerk and Coates (yes him of Crash-2 – top bloke), looked at patients with sepsis in the ED. They measured StO2 on arrival, on leaving and at 24 hours post arrival. The abstract is worth a read so do that now.
So there you go. 49 patients with sepsis were included in the trial and they demonstrated that at baseline (ED admission) there was no difference in St02 between survivors and non-survivors. However, by the time the patients left the ED the St02 in survivors had changed significantly and this was still a difference at 24 hours. All very interesting I hear you say, but let us delve a little deeper…..
“Who is studied in this paper?”] Good question. We’ve looked at a number of papers around sepsis for some time and we do find it interesting that the definitions for what is sepsis and what is septic shock can vary. In this paper it’s a little unclear what the definition is, but it’s really important to know. If it’s just SIRS criteria then there are few patients with tonsillitis who might get into the study (I don’t think this was the case), but clear boundaries are important.
The other thing that we picked up on when looking at this study was that the patient groups were just a bit too tidy. Everyone in the study had sepsis…but I don’t know about you, I sometimes find that the initial diagnosis is wrong. Patients in the analysis were excluded if their shock was due to another reason (e.g. PE) and it just begs the question about whether they were excluded before or after data collection. The flow diagram appears to show that the other cases were excluded after screening and before testing. If so that’s OK, but…, it still seems very tidy.
Ideally I’d like to see a study where patients are analysed on an intention to analyse basis (like intention to treat), so we take all comers that we thought had sepsis at the beginning rather than the ones we know had sepsis at the end. It makes it more real.
“Is StO2 an independent marker of risk of death?”] That is sort of a key question really. There are arguably two reasons to use this non-invasive test. Firstly it might be really easy to do. Secondly it might convey an advantage in terms of care. Now in terms of the first it’s a fairly reasonable no-brainer in that you just put a pad on the thumb and turn the machine on. With regard to the second then it’s far more complex.
In this study the usefulness cannot be measured as the clinicians were blinded to the readings so there is no treatment effect. However, we should consider whether this might offer an advantage over what we already have..so what’s that? Well, for us it’s lactate (venous or arterial) as a marker of tissue perfusion. We use it a lot in our ED as we have a co-located ABG machine, and as Alan has taught us VBGs are fab. So if we already have a measure of tissue perfusion is this any better or different? I guess that’s the question that remains unanswered here. In the paper there are clear differences at baseline between the two groups in terms of illness severity (as defined by presence of septic shock) and lactate, which in the survivor group at baseline is markedly different to the non-survivors (survivors baseline = 3.7 (2.8 to 4.5) non-survivors =6.3 (4.5 to 8.1)).
So I am still confused as to whether this may offer any advantage over what we already have. At the moment it would appear that StO2 tells me things which I already know. I’d have loved to see whether it is an independent marker for mortality, but that would require a logistic regresssion analysis and the numbers are just far too small here for that to be done. The paper cannot answer this question and I am therefore left with the feeling that StO2 and lactate might actually be measuring the same thing – and I do lactate already so what’s the potential benefit?
“This study is all about sepsis, what about other conditions?”] Obviously this paper looks only at septic patients and I’ve given it a hard time as I think we already have alternatives in the resus room. However, out on the road, in the air, on a boat things are different and perhaps this is something that might be more applicable in a pre-hospital setting. Interestingly this has been looked at by Lyon, Hutchinson and Lockey in a feasibility study of 10 patients. Small stuff I agree, but they suggest that it might be feasible and offer some advantages, but only if the kit gets smaller and lighter.
Guyette et al looked at 150 cases in the US and found links to low StO2 and need for life saving interventions, but I’m not sure that that’s a surprise.
Interesting thoughts though. If we started out from the premise that simple cardiovascular markers are a bit rubbish then this is an areas where I would like to see more research.
“Is this the right sort of study to answer my question?” Hmm, well I guess that depends on what your question was! If it was ‘are patients with an unchanging StO2 more likely to die in sepsis’ then the answer might be yes. A blinded prospective diagnostic cohort can answer this question. What it cannot tell you is whether it is going to be an independent predictor, nor if it is an independent predictor if it is going to be useful for patient care. That would require an RCT and I have not been able to find one in emergency medicine that comes close to helping us answer the question.
“Should I buy an StO2 monitor now?” Er, that’s up to you but I’m not buying one just yet, unless someone wants to fund me to do the research that needs to be done.
Who is using this in the ED at the moment?
No idea. Would love to hear from anyone who is though. Specifically want to know if it gives them more info than a blood gas.
Before you go please don’t forget to…
- Subscribe to the blog (look top right for the link)
- Subscribe to our PODCAST on iTunes
- Follow us on twitter @stemlyns
- Like us on Facebook
- Find out more about the St.Emlyn’s team