Following Paul Myles’ paper in Anesthesiology in 2007 – that demonstrated bad outcomes in patients anaesthetised with nitrous oxide (click here), “experts” clamoured to demand that we stop using the stuff in our clinical practice. Their opinions were enhanced by the ENIGMA trial, that claimed increased risk of myocardial infarction in patients receiving nitrous (click here); following adjustment for the usual factors. I have been personally accused of “poisoning” my patients by continuing to administer nitrous. Hence, it was with great relief that I read this paper (click here) in this month’s anesthesia and analgesia.
Turan and Colleagues evaluated almost 50,000 patients who had noncardiac surgery at the Cleveland Clinic over a 4 year period (2005 and 2009). Of the patients that had general anesthesia, 17,00 were given N2O (45%) and 21,000 were not (55%). Of each group, 10,000 patients were propensity score-matched on 30-day mortality and a set of 8 in-hospital morbidity/mortality outcomes.
The results were surprising. Patients that were given N2O intraoperatively had decreased odds of 30-day mortality (odds ratio [OR]: 97.5% confidence interval, 0.67, 0.46–0.97; P= 0.02), compared with no nitrous. In addition, patients that received had a17% (OR: 0.83, 0.74–0.92) reduced odds of experiencing major in-hospital morbidity/mortality than non-nitrous (P < 0.001). In particular, the risk of pulmonary complications with significantly lower in patients who received nitrous.(OR, 95% Bonferroni-adjusted CI: 0.59, 0.44–0.78).
Ok – so this was a propensity score analysis induced fluke – right? In the same issue of A&A we have a second paper that analysed the POISE trial outcomes (click here). 30% of the 6000 patients in the study received nitrous – and there was NO association between the gas and adverse outcomes. A fairly biased editorial in A&A, written with the help of Paul Myles, whose group is the only one that has demonstrated bad outcomes with nitrous, dismembers the Turan paper.
Nitrous oxide has been around for 160 years. I am not aware that there is a pandemic of death and MI amongst the patients of those of us who use the stuff. In any case, I think that this paper, at the very least, suggests that the jury is still out on the subject.
AnaesthesiaWest
WESTERN ANAESTHESIA SOCIETY IRELAND
AnaesthesiaWest 
As an avowed opponent of the blue poison i thought I’d make a few comments about this.
Before mentioning this study i think it’s necessary to compare this study with the ENIGMA study. Enigma was a large, multi-centre, multinational randomised controlled trial of over 2000 patients. It is true that it is the only RCT that has shown bad outcomes with N20 but that is because it is the only trial that has ever looked. It doesn’t matter how long nitrous has been around, if you don’t look for something you won’t see it. Also, this was done in multiple hospitals in Australia and Hong Kong and is not a single center study as implied above or as is the Turan study.
It has been criticised due to the fact that some patients in the nitrous-free group mainly got 80% oxygen – confounding the effect of n20 with that of high inspired oxygen concentration. The authors did look into the effect afterwards and found no association with wound infection and fi02, however this is still a potential issue.
Enigma showed a reduction in wound infection, severe nausea, atelectasis and pneumonia in the nitrous free group, as well as a non significant trend to reduced MI. In view of recent studies showing a higher death rate with higher Fi02 one could speculate that the absence of a difference in mortality despite a higher fi02, in a population with a large number having cancer surgery might have further masked a beneficial effect of avoiding nitrous.
Contrast this with an observational study, done by people who KNOW that nitrous may make people have more complications, and may therefore “treat” high risk patients by avoiding nitrous, creating a higher risk group. As noted in the editorial to the Turan study, for observational studies being large doesn’t make it better and may in fact make it worse. The propensity scoring can help but you can only control for variables and confounders you know about.
That said there is definitely some hope for nitrous lovers with this subgroup analysis from the ENIGMA trial, showing less chronic pain in those who had nitrous- quite plausible given nitrous’s NMDA antagonist effects.
http://www.ncbi.nlm.nih.gov/pubmed/21889262
It will be interesting to see the results of ENIGMA-II. This is an even larger (7000 patients) trial of nitrous oxide, powered to detect differences in myocardial infarction rates. I was involved in recruiting patients for this in 2009 so it can’t be too far off coming out. Pending the results of this and in view of the chronic pain results as much as i hate the blue poison it’s use is still more then reasonable.
It may turn out that the current situation where people always give or always avoid nitrous is wrong and we need a patient specific approach. Avoid it for those at risk of MI or respiratory complications and give it to those at risk of chronic pain.