You could say that we might have a bit of a thing for nitrous oxide here at Rapid Sequence. We've looked it a few times before (Nitrous & Nausea, No Place For Nitrous Anymore) and once again we’re back for some more analysis. I suppose this may reflect the fact that nitrous is one of those aspects of anaesthesia that is a bit of a Marmite, polarising anaesthetists over its benefits vs lethality and thus providing ample fodder for discussion (occasionally evidence based). Today’s blog looks at what had the potential to be a pretty big milestone in nitrous’ long illustrious history; the ENIGMA II trial. We got a little bit excited about these results a few months ago after the authors released a little teaser trailer, but now the results are finally here in full (actually published in August but FRCA exams etc etc). This was the highly anticipated sequel to the somewhat flawed (but interesting) initial ENIGMA trial that raised questions about the link between nitrous and post-op cardiovascular events. Would this add more nails to nitrous’ coffin lid or would it breathe new life into this veteran anaesthetic drug? If you haven't read the article yet, check it out here:
The Lancet; The Safety of Addition of Nitrous Oxide to General Anaesthesia in At-Risk Having Major Non-Cardiac Surgery
The Lancet; The Safety of Addition of Nitrous Oxide to General Anaesthesia in At-Risk Having Major Non-Cardiac Surgery
What's It About?
If you haven’t read much about the ENIGMA trial (Elimination of Nitrous from the Gas Mixture in Anaesthesia) and its follow up analyses, it raised a few (not particularly well supported) questions about nitrous’ safety by suggesting a link between the use of nitrous and cardiovascular events post-operatively. The effects of nitrous oxide in elevating plasma homocysteinaemia levels (a risk factor for cardiovascular events in other scenarios) seems fairly clear, but the clinical significance of it does not. The flaws in the ENIGMA trial therefore led to the conception of ENIGMA II to try and answer these questions. Put most simply, does nitrous kill you?
So What've They Done?
If you compare it to the initial ENIGMA, they’ve basically gone bigger and better. Compared to ENIGMA’s 2050 patients in a trial that was, by their own admission, not really designed or powered to assess cardiovascular complications, ENIGMA II recruited 7112 patients of medium to high cardiovascular risk in a randomised single-blinded study that was specifically designed to look and these outcomes as well as respond to the criticisms of the initial trial.
Patients who had cardiovascular risk factors (e.g. known IHD or cerebrovascular disease), and were undergoing non cardiac surgery of longer than 2 hours were elligile. They were randomised to receive either 70% nitrous oxide with 30% O2 or an air-oxygen mix with 30% O2. The anaesthetist was aware of the allocation but no-one else in the process was (patient, surgical team, research team). The patients had close follow up whilst in hospital, including specifically timed ECG and troponin investigations, and telephone follow up at 30 days. The primary outcome was a composite of death and cardiovascular event, with a number of secondary and tertiary outcomes, including surgical site infection, post-op nausea and vomiting (PONV), all-cause mortality, and ICU admission.
Patients who had cardiovascular risk factors (e.g. known IHD or cerebrovascular disease), and were undergoing non cardiac surgery of longer than 2 hours were elligile. They were randomised to receive either 70% nitrous oxide with 30% O2 or an air-oxygen mix with 30% O2. The anaesthetist was aware of the allocation but no-one else in the process was (patient, surgical team, research team). The patients had close follow up whilst in hospital, including specifically timed ECG and troponin investigations, and telephone follow up at 30 days. The primary outcome was a composite of death and cardiovascular event, with a number of secondary and tertiary outcomes, including surgical site infection, post-op nausea and vomiting (PONV), all-cause mortality, and ICU admission.
What Did They Find?
In short, nothing! There was no difference in the main outcomes between the two groups (except a higher incidence of PONV in the nitrous group, unsurprisingly). After randomly allocating 7112 patients, 7011 were included in the study at baseline and 6992 were successfully followed up for the primary endpoint (follow up rate over 99.7% !). The primary endpoint occurred in 579 (8%) of patients, with 283 in the nitrous group and 296 in the non-nitrous group. Looking at some of the secondary outcomes, surgical site infections occurred in 321 (9%) of the nitrous group and 311 (9%) of the non-nitrous group, whilst severe PONV occurred in 506 (15%) of the nitrous group compared to 378 (11%) of the non-nitrous group.
Is It Any Good?
Overall I would definitely have to say yes. There are a lot of good points to mention when picking through this trial and this adds a bit of weight to its conclusions. I think a lot of this has come from building on the flaws of the initial ENIGMA trial making the necessary improvements in study design. To start with it’s a good sized randomised (single) blinded trial that knows exactly what it is looking for. They are clear with their power calculations to detect what seems to be a sensible degree of difference (a change in the incidence of the primary outcome from 6 to 8 % with the standard values for type 1 and 2 errors as 0.05 and 0.1 respectively). The approach to selecting high risk patients for the trial would appear to promote its sensitivity in detecting the clinical outcomes that they are looking for, and the very high follow up rate of 99.7% removes that uncertainty of missing data perhaps playing a role in the results.
With regards to measurement and observation, they have been clear in their definitions of the outcomes they are looking at; a potential hazard when describing myocardial ischaemia/infarction. The data showing the different outcomes and other variables of the two groups (ASA status, baseline meds, types of operation, anaesthetic drugs used, etc.) are clearly presented and similar between the groups, as would be expected from randomisation.
Where do they fall down then? To be honest the criticisms of the study are more nit-picking in nature than strongly questioning the validity of the study’s conclusions. I suppose the first question that came to my mind was that they have been selective about the population they wanted to look at; namely the patients with the higher cardiovascular risk. This is done with the reasoning that it would make it easier to detect the desired outcome in this population, because these are the ones that are most at risk of post-operative events. But this is assuming that these will also be the patients that are most at risk from nitrous’ side effects, and that we can then extrapolate the results back to the lower risk population. Not an unreasonable preposition but still involves some extrapolation of the results. Secondly, they lose some marks for the absence of double-blinding. It is certainly impractical and probably impossible to safely blind anaesthetists to the gas mixture they are given, but it does bring in the risk that the anaesthetist knowing which gas mixture they are giving might affect the other aspects of the anaesthetic they give. The authors have approached this problem by looking at the drugs given and intraoperative variables, noting a few small differences like the incidence of antiemetic prophylaxis being given, but there is always the potential to miss things and hence why double blinding is the ideal scenario.
With regards to measurement and observation, they have been clear in their definitions of the outcomes they are looking at; a potential hazard when describing myocardial ischaemia/infarction. The data showing the different outcomes and other variables of the two groups (ASA status, baseline meds, types of operation, anaesthetic drugs used, etc.) are clearly presented and similar between the groups, as would be expected from randomisation.
Where do they fall down then? To be honest the criticisms of the study are more nit-picking in nature than strongly questioning the validity of the study’s conclusions. I suppose the first question that came to my mind was that they have been selective about the population they wanted to look at; namely the patients with the higher cardiovascular risk. This is done with the reasoning that it would make it easier to detect the desired outcome in this population, because these are the ones that are most at risk of post-operative events. But this is assuming that these will also be the patients that are most at risk from nitrous’ side effects, and that we can then extrapolate the results back to the lower risk population. Not an unreasonable preposition but still involves some extrapolation of the results. Secondly, they lose some marks for the absence of double-blinding. It is certainly impractical and probably impossible to safely blind anaesthetists to the gas mixture they are given, but it does bring in the risk that the anaesthetist knowing which gas mixture they are giving might affect the other aspects of the anaesthetic they give. The authors have approached this problem by looking at the drugs given and intraoperative variables, noting a few small differences like the incidence of antiemetic prophylaxis being given, but there is always the potential to miss things and hence why double blinding is the ideal scenario.
Final Thoughts
For me this adds more information to debate on the pros and cons of nitrous oxide, strongly suggesting that increased cardiovascular event risk is not a feature of nitrous use, a con that may have weighed on some practitioner’s minds after ENIGMA (and others). They have also tried to shed more light on the risk of surgical site infections, another area that had been questioned in the past, and the results are again reassuring for us. That said, this study again demonstrates a link between nitrous use and severe PONV. With these operations being at least 2 hours in duration I suppose this isn’t surprising (check out our recent blog on nitrous duration and PONV), but this may be enough of a downside for some clinicians that they will still avoid using it.
Ultimately, I think this study does a lot for nitrous’ safety profile. As has been commented in some of the reviews in the past, nitrous has a long long history, and you don’t usually hang around that long if you’re a bit of a killer. It is a drug, it has side effects and it has desired effects. The side effects here are probably less worrying than we had been starting to think, so we are left to balance them against the plus points to provide the best anaesthetic for our patients. I can see scenarios where I would choose to use nitrous and others where I would like to avoid it. Now I can have more confidence that I am not dishing out myocardial infarctions when I do turn it on. I just might need to grab some extra ondansetron though.
Once again thanks for reading, and please let me know your thoughts on the subject. What role does nitrous play in your anaesthetics and has ENIGMA II changed this at all?
Tom Heaton
Ultimately, I think this study does a lot for nitrous’ safety profile. As has been commented in some of the reviews in the past, nitrous has a long long history, and you don’t usually hang around that long if you’re a bit of a killer. It is a drug, it has side effects and it has desired effects. The side effects here are probably less worrying than we had been starting to think, so we are left to balance them against the plus points to provide the best anaesthetic for our patients. I can see scenarios where I would choose to use nitrous and others where I would like to avoid it. Now I can have more confidence that I am not dishing out myocardial infarctions when I do turn it on. I just might need to grab some extra ondansetron though.
Once again thanks for reading, and please let me know your thoughts on the subject. What role does nitrous play in your anaesthetics and has ENIGMA II changed this at all?
Tom Heaton
References
- The Lancet; The Safety of Addition of Nitrous Oxide to General Anaesthesia in At-Risk Having Major Non-Cardiac Surgery. 2014
- Anesthesiology: Avoidance of Nitrous Oxide for Patients Undergoing Major Surgery: A Randomized Controlled Trial. 2008
- Anesthesia & Analgesia: Nitrous oxide and long-term morbidity and mortality in the ENIGMA trial. 2011
Image courtesy of David Castillo Dominici/freedigitalphotos.net
RSS Feed
