We use a lot of NIV in Virchester. It’s a great technique that can be started in the emergency department as a supportive therapy or as a bridge to intubation and ICU care. The machines we use (mostly the Carina) are easy to set up, easy to read and easy to adjust to the patient. It sounds easy, and a lot of the time it is, but there are still issues. Although the technology is great, the interface with the patient has its problems. We use face masks (in three sizes) to connect the machine to the patient and whilst these are clearly well designed, for a lot of our patients it’s a real struggle to get a seal. The edentulous, the bearded and those patients with anything other than standard craniofacial anatomy present a challenge in terms of getting a tight enough seal for the system to maintain good airway pressure and support. In the short term (minutes) then some NP airways and a really tight pull on the straps will usually get us there, but it’s not satisfactory for prolonged use with the risk of intolerance, pressure sores and pain.
I have a lot of sympathy for the patients. At a time when they feel really short of breath the clinicians (us) put a tight fitting mask over their face. You can understand how this might feel counterintuitive to those fighting for breath.
There are alternatives of course; pressurised hoods have been around for some time and in the Virchester journal club we’ve reviewed a few papers over the years looking at their effectiveness for weaning patients from ventilation.
This month I found an interesting RCT from JAMA looking at the use of hoods vs. masks in a medical ICU to prevent the need for intubation (thus potentially avoiding many of the known hazards of mechanical ventilation). As always make sure that you read the full paper (paywalled), and then read on.
It’s an RCT, and since we are considering a trial of a therapy it is the appropriate method for this kind of study. Patients studied were those on the ICU with a diagnosis of ARDS (according to Berlin criteria). Subjects were randomised to either face mask NIV or to support using a hood device with a soft seal around the neck. Patients with severe disease or who were unable to participate due to anatomical or ceiling of care reasons were not included. Usually we want to see blinding in RCTs but of course in a trial like this that’s simply not possible. Allocation concealment is vital in trials like this – they randomised by computer generated permuted block – great – but then left this sequence in the dreaded sealed opaque envelope. Are we sure people didn’t tamper or curiosity wasn’t peaked?
This study was designed as a single centre trial with an initial target of recruiting over 200 patients. However, they only recruited 83 patients into the trial as the data monitoring committee stopped the trial early when they felt that it was clear that there was a benefit to the helmet therapy.
As an aside Data Monitoring Commmittees are really interesting aspects of clinical trials. They perform pre-planned interim analyses on data to ensure that if a point is reached where there is no point in carrying on, then it’s time to stop. I sit on a DMC for a UK trial and the roles and responsibilities are considerable. The DMC usually only stops a trial early when there is clear evidence of success or harm with one of the treatments.
What are the main results?
In this study the main outcome measure was the rate of intubation and the outcomes in the two groups were really very different. In the helmet group 18.2% of patients ended up being intubated as compared to 61.5% in the face mask group. This was statistically significant and of course the clinical difference is huge. It equates to an absolute risk reduction of 43.3% or a number needed to treat (NNT) of 2.3. If that’s reproducible in normal practice that’s a really important clinical intervention.
They also looked at mortality both in-hospital and at 90 days. Both of these were much better in the helmet group (34.1% vs. 56.4%). Again that’s a huge difference. However, there should be a bit of caution here – the idea that using a different modality to deliver what is essentially the same intervention (NIV at set pressures) can reduce mortality in ARDS by more than any other previous intervention (including proning, paralysis, and lung protective ventilation) doesn’t really have face validity.
Any concerns with this study?
I think we all need to be cautious when looking at single centre studies with large treatment effects. There is a long history of the results of such trials being less dramatic when attempts are made at reproducing them elsewhere (an challenge pitched at the River’s sepsis trials for example). To be fair the authors acknowledge this and ask for a larger multicentre trial and this is clearly a wise suggestion.
Only 10% of eligible patients were recruited thus limiting the generalisability to the ED/ICU cohort.
There are some interesting differences in the baseline populations with potentially clinically significant differences in the number of patients with solid organ malignancy, pneumonia, aspiration pneumonia and extra pulmonary ARDS.
More patients were shocked in the facemask group and were 5 times more were on vasopressors at recruitment.
The helmet group had a higher median delivered PEEP (8 vs 5). Whilst this may reflect tolerability and an advantage of the helmet system, it also suggests the patients had slightly different therapy post randomisation.
The critieria for deciding when to intubate were also unclear and left to the discretion of the clinicians. Although this is pragmatic (good) it does increase the chance of bias, especially in a trial that is unblinded. If the clinicians believed in helmets more than facemasks they might be more inclined to persist with the new therapy. We don’t know if this happened, but it is certainly a risk.
All of this suggests to me that the facemask group were in fact different and potentially much sicker and likely to have poorer outcomes from the start. This is unfortunately what happens when you randomise small numbers – baseline confounders are not evenly distributed and treatment effects can be markedly exaggerated.
What does this mean for the ED?
Looking at the patients in this trial then it’s clear that they are very different from those in my department. The high number of patients with malignancy, the ICU setting and the fact that this was undertake in a single centre make transferring this information directly into the resus room very difficult.
Back in resus I’m still struggling with my patient on NIV with a bit of agitation and a poorly fitting mask. Should I ask for a helmet? Well, currently I don’t have one so the answer is easy – not at the moment. As for the future then perhaps, the evidence of effectiveness in this trial is impressive, but I would need to see data from other centres and from a population based in the ED before I change my practice. Remember that we do have an alternative therapy for failed NIV at the moment; I can still intubate and ventilate my ED patients and so for now that will continue to be my strategy. Perhaps this paper will serve as a call to arms to trial helmet NIV in the ED and that could be a really interesting RCT.
As always I’d be really interested to hear from you and especially if this is a therapy you’re using or thinking of using in the ED as opposed to the critical care setting. I’d also love to see the Bottom Line team have a look at this from the CritCare perspective.
Simon and Dan
Berlin criteria for ARDS. http://lifeinthefastlane.com/ccc/acute-respiratory-distress-syndrome-ards-definitions/
Data monitoring committees in clinical research. https://en.wikipedia.org/wiki/Data_monitoring_committee
Oronasal mask versus helmet in acute hypercapnic respiratory failure. Eur Respir J. 2015 Mar;45(3):691-9. doi: 10.1183/09031936.00053814. Epub 2014 Dec 10.
Effect of Noninvasive Ventilation Delivered by Helmet vs Face Mask on the Rate of Endotracheal Intubation in Patients With Acute Respiratory Distress Syndrome: A Randomized Clinical Trial. JAMA. 2016 Jun 14;315(22):2435-41. doi: 10.1001/jama.2016.6338.
High-Flow Nasal Oxygen vs Noninvasive Positive Airway Pressure in Hypoxemic Patients After Cardiothoracic Surgery: A Randomized Clinical Trial. JAMA. 2015 Jun 16;313(23):2331-9. doi: 10.1001/jama.2015.5213.
Noninvasive Ventilation for the Emergency Physician.Emerg Med Clin North Am. 2016 Feb;34(1):51-62. doi: 10.1016/j.emc.2015.08.004.
Noninvasive Ventilation and the critically ill. http://lifeinthefastlane.com/non-invasive-ventilation/
Pre-hospital non-invasive ventilation for acute respiratory failure: a systematic review and cost-effectiveness evaluation. Health Technol Assess. 2015 Jun;19(42):v-vi, 1-102. doi: 10.3310/hta19420.
Non-invasive ventilation and continuous positive pressure ventilation in emergency departments: where are we now? Emerg Med J. 2008 Apr;25(4):190-4. doi: 10.1136/emj.2007.049072.
Before you go please don’t forget to…