This blog reviews the first large randomised controlled trial of cricoid pressure in emergency patients as published in JAMA Surgery, October 2018.
Cricoid pressure is controversial. Although it has been taught in anaesthesia practice, since it’s original description by Sellick1, as an essential requirement when performing RSI, the evidence for its benefit and lack of risk is far from compelling2,3. The late Dr John Hinds famously brought this to the attention of the #FOAMed world in his ‘Cricolol’ presentation concluding that the evidence for its effectiveness was very poor indeed. If you’ve never seen that presentation then you can watch it on the link below from EMCRIT.
John’s presentation was hilarious but even he would admit that the major problem was a lack of definitive evidence. This month we have the publication of the larget trial to date on the use of cricoid pressure in emergency patients4. You can read the paper here and it appears to be open access at the moment. Read the abstract below, but don’t forget to read the full paper online.
What kind of study is this?
Refreshingly it’s a randomised controlled trial. As cricoid is an intervention then it is entirely appropriate design. Interestingly it is a non-inferiority trial of which we will lean more later. In essence the authors have designed a trial not to demonstrate a difference between cricoid and no cricoid, but to ask whether a lack of cricoid is inferior to the use of cricoid pressure.
Tell me about the patients
This is important as the risks associated with aspiration on the induction of anaesthesia vary depending on the patient group. In this study they looked at operating room patients who were at potential risk of aspiration for reasons such as obesity, hiatus hernia, lack of fasting etc. In this respect they were similar to ED patients but not entirely, the ED patient population is arguably even more risky. The point is that this was an at risk group and therefore not unreasonable as a group to study. Only adults were included and the pregnant population were excluded.
They recruited 3472 patients in 10 French hospitals between 2014 and 2017.
What about the intervention?
Basically patients were randomised to cricoid pressure or to sham pressure where the operator put their hands in position but did not press down. Paralysis was almost always achieved with suxamethonium, but other elements of induction such as opiates and induction agent were left to the discretion of the anaesthetist.
The primary outcome was evidence of aspiration either at the time of laryngoscopy or on subsequent tracheal aspiration.
What about the results?
The primary outcome of aspiration was no different in the two groups; 0.6% in the cricoid group and 0.5% in the sham group. On the authors statistical analysis they argue that they failed to demonstrate the noninferiority of the sham procedure. However, the trial recruitment targets were based on an estimated prevalence of aspiration of 2.8% which is far higher. The effect of this lower prevalence in the study means that the study is effectively underpowered to detect the inferiority margins (a 50% difference) that the authors sought. Indeed the potential differences the authors were considering were actually quite large even with a 2.8% prediction of aspiration rate.
So we have ended up with an underpowered trial and thus it is tempting to dismiss the results, but perhaps we should think again. The results tell us that the incidence of aspiration is actually very low, and that there does not appear to be any signal that cricoid affects the incidence of aspiration. If you are interested in this sort of thing you can use an online stats calculator to work out what the sample size required would be for a study that assumed an aspiration rate of 0.5% and which sought to investigate a non-inferiority rate of 1.5x the expected (a difference they sought in this study). I’ll save you the trouble and tell you that my estimate is that it’s roughly a trial recruiting about 34,000 patients, so 9-10 times the number of patients in this study5.
If you are feeling particularly nerdy then you might ask yourself what the confidence intervals around the rate of aspiration are across the whole cohort. This is a bit naughty as a post hoc thought, but if we consider the overall rate of aspiration in this study to be 19 patients in 3472 studied then the event rate overall is 0.55% and the confidence intervals go from 0.3% to 0.8%. So basically this is a pretty rare event and thus any intervention that you are going to use to affect the overall rate is going to have a very large NNT unless it is incredibly effective, which on the results of this trial seem highly unlikely.
That’s the primary outcome, but what about the secondary outcomes? They also looked at difficulty, laryngoscopy and intubation times and in this respect there were arguably advantages to the sham technique, but the importance of these clinically is probably not that great.
You should also look at other reviews of this paper that place the results in context.
- JAMA Surgery Editorial by Samuel Tisherman6
- The case of the inferior superiority on EMCRIT with Rory Spiegal7
What’s the clinical bottom line?
On the basis of this and other trials there is no compelling evidence for the use of cricoid pressure in operating department patients, but there is no real evidence of serious harm either. It looks likely that laryngoscopy may be easier without it in place.
So on balance, and arguably not on the strongest of evidence cricoid pressure is probably not doing a great deal in this group of patients. However, there are groups of patients in whom this evidence clearly does not apply, the pregnant, the paediatric, and perhaps those in whom regurgitation is extremely likely such as small bowel obstruction. In those cases I’ll take the judgement of an expert anaethetist every time. We must always be cautious in extending the evidence beyond the patient populations it was created with.
Specifically for us in the ED we don’t know whether we can directly transfer that opinion yo our patients. However, other studies, albeit smaller and less rigorous ones also question the effectiveness of cricoid in emergency patients8–11. In Virchester I’ll carry on not using it in most cases, but I accept that this is an evidence light decision.