Nitrous oxide‐based techniques versus nitrous oxide‐free techniques for general anaesthesia

Rao Sun; Wen Qin Jia; Peng Zhang; KeHu Yang; Jin Hui Tian; Bin Ma; Yali Liu; Run H Jia; Xiao F Luo; Akira Kuriyama

doi:10.1002/14651858.CD008984.pub2

. 2015 Nov 6;2015(11):CD008984. doi: 10.1002/14651858.CD008984.pub2

Nitrous oxide‐based techniques versus nitrous oxide‐free techniques for general anaesthesia

Rao Sun ¹, Wen Qin Jia ², Peng Zhang ³, KeHu Yang ^4,^✉, Jin Hui Tian ¹, Bin Ma ¹, Yali Liu ¹, Run H Jia ⁵, Xiao F Luo ⁶, Akira Kuriyama ⁷

Editor: Cochrane Anaesthesia Group

PMCID: PMC9326975 PMID: 26545294

Abstract

Background

Nitrous oxide has been used for over 160 years for the induction and maintenance of general anaesthesia. It has been used as a sole agent but is most often employed as part of a technique using other anaesthetic gases, intravenous agents, or both. Its low tissue solubility (and therefore rapid kinetics), low cost, and low rate of cardiorespiratory complications have made nitrous oxide by far the most commonly used general anaesthetic. The accumulating evidence regarding adverse effects of nitrous oxide administration has led many anaesthetists to question its continued routine use in a variety of operating room settings. Adverse events may result from both the biological actions of nitrous oxide and the fact that to deliver an effective dose, nitrous oxide, which is a relatively weak anaesthetic agent, needs to be given in high concentrations that restrict oxygen delivery (for example, a common mixture is 30% oxygen with 70% nitrous oxide). As well as the risk of low blood oxygen levels, concerns have also been raised regarding the risk of compromising the immune system, impaired cognition, postoperative cardiovascular complications, bowel obstruction from distention, and possible respiratory compromise.

Objectives

To determine if nitrous oxide‐based anaesthesia results in similar outcomes to nitrous oxide‐free anaesthesia in adults undergoing surgery.

Search methods

We searched the Cochrane Central Register of Controlled Trials (CENTRAL; 2014 Issue 10); MEDLINE (1966 to 17 October 2014); EMBASE (1974 to 17 October 2014); and ISI Web of Science (1974 to 17 October 2014). We also searched the reference lists of relevant articles, conference proceedings, and ongoing trials up to 17 October 2014 on specific websites (http://clinicaltrials.gov/, http://controlled‐trials.com/, and http://www.centerwatch.com).

Selection criteria

We included randomized controlled trials (RCTs) comparing general anaesthesia where nitrous oxide was part of the anaesthetic technique used for the induction or maintenance of general anaesthesia (or both) with any general anaesthesia using a volatile anaesthetic or propofol‐based maintenance of anaesthesia but no nitrous oxide for adults undergoing surgery. Our primary outcome was inhospital case fatality rate. Secondary outcomes were complications and length of stay.

Data collection and analysis

Two review authors independently assessed trial quality and extracted the outcome data. We used meta‐analysis for data synthesis. Heterogeneity was examined with the Chi² test and by calculating the I² statistic. We used a fixed‐effect model if the measure of inconsistency was low for all comparisons (I² statistic < 50%); otherwise we used a random‐effects model for measures with high inconsistency. We undertook subgroup analyses to explore inconsistency and sensitivity analyses to evaluate whether the results were robust. We assessed the quality of evidence of the main outcomes using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) system.

Main results

We included 35 trials (13,872 adult participants). Seven included studies were at low risk of bias. We identified eight studies as awaiting classification since we could not obtain the full texts, and had insufficient information to include or exclude them. We included data from 24 trials for quantitative synthesis. The results of meta‐analyses showed that nitrous oxide‐based techniques increased the incidence of pulmonary atelectasis (odds ratio (OR) 1.57, 95% confidence interval (CI) 1.18 to 2.10, P = 0.002), but had no effects on the inhospital case fatality rate, the incidence of pneumonia, myocardial infarction, stroke, severe nausea and vomiting, venous thromboembolism, wound infection, or the length of hospital stay. The sensitivity analyses suggested that the results of the meta‐analyses were all robust except for the outcomes of pneumonia, and severe nausea and vomiting. Two trials reported length of intensive care unit (ICU) stay but the data were skewed so were not pooled. Both trials reported that nitrous oxide‐based techniques had no effects on the length of ICU stay. We rated the quality of evidence for two outcomes (pulmonary atelectasis, myocardial infarction) as high, four outcomes (inhospital case fatality rate, stroke, venous thromboembolism, length of hospital stay) as moderate, and three (pneumonia, severe nausea and vomiting, wound infection rate) as low.

Authors' conclusions

Given the evidence from this Cochrane review, the avoidance of nitrous oxide may be reasonable in participants with pre‐existing poor pulmonary function or at high risk of postoperative nausea and vomiting. Since there are eight studies awaiting classification, selection bias may exist in our systematic review.

Plain language summary

Nitrous oxide (laughing gas)‐based techniques versus nitrous oxide‐free techniques for general anaesthesia

Review question

We reviewed the evidence about the harmful effects of nitrous oxide on people undergoing general anaesthesia.

Background

Nitrous oxide is an anaesthetic gas which has been used for more than 160 years for inducing anaesthesia and keeping patients anaesthetized throughout an operation. It is also known as 'laughing gas'. It is a colourless non‐flammable gas with a pleasant, faint sweet odour and taste. Its low cost and low toxicity have made nitrous oxide by far the most commonly used general anaesthetic. However, some studies have reported that adding nitrous oxide may lead to harmful effects. This has led many anaesthetists to question its continued routine use in a variety of operating room settings.

We wanted to discover whether using nitrous oxide in general anaesthesia was better or worse than not using nitrous oxide.

Study characteristics

We examined the evidence available up to 17 October 2014. We included 35 trials involving 13,872 adult participants, all of whom were randomized to either receive nitrous oxide or no nitrous oxide. The trials covered a variety of situations during general anaesthesia.

Key results

We found that general anaesthesia with nitrous oxide increased the risk of pulmonary atelectasis (i.e. failure of the lungs to expand fully). When we restricted the results to the highest quality studies only, we found evidence that nitrous oxide may potentially increase the risk of pneumonia and severe nausea and vomiting. However, nitrous oxide had no effect on the patients' survival, the incidence of heart attack, stroke, wound infection, the occurrence of blood clots within veins, the length of hospital stay, or the length of intensive care unit stay.

Quality of the evidence

The evidence related to survival of participants was of moderate quality because we did not have enough data. The evidence related to some harmful effects, such as failure of the lungs to expand fully and heart attack, was of high quality, while for other harmful effects, such as stroke and the occurrence of blood clots within veins, the evidence was of moderate quality. For others, such as pneumonia, severe nausea and vomiting, and wound infection, the evidence was of low quality. The evidence related to the length of time spend in hospital was of moderate quality.

Authors conclusions

The avoidance of nitrous oxide may be reasonable in participants with pre‐existing poor pulmonary function or at high risk of postoperative nausea and vomiting.

Summary of findings

Background

Description of the condition

Nitrous oxide, also known as laughing gas, is a colourless non‐flammable gas with a pleasant, faintly sweet odour and taste. The gas has been in use for more than 160 years for the induction and maintenance of general anaesthesia. It has been used as a sole agent but is most often employed as part of a technique using other anaesthetic gases, intravenous agents, or both. Its low tissue solubility (and therefore rapid kinetics), low cost, and low rate of cardiorespiratory complications have made nitrous oxide by far the most commonly used general anaesthetic. Worldwide, it is given to more than one billion surgical patients annually (Fleischmann 2005).

The accumulating evidence regarding adverse effects of nitrous oxide administration has led many anaesthetists to question its continued routine use in a variety of operating room settings. Adverse events may result from both the biological actions of nitrous oxide and the fact that to deliver an effective dose, nitrous oxide, which is a relatively weak anaesthetic agent, needs to be given in high concentrations that restrict oxygen delivery (for example, a common mixture is 30% oxygen with 70% nitrous oxide).

The disadvantages of nitrous oxide have been reported. Concerns have been raised regarding the risk of compromising the immune system (Parbrook 1967), low blood oxygen levels (Cheney 2007), impaired cognition (mental ability) (Culley 2007; Linde 1969), postoperative cardiovascular complications (Myles 2008b), as well as bowel obstruction from distention and possible respiratory compromise (Eger 1965). In addition, nitrous oxide may increase the risk of developing brain damage from reduced cerebral blood flow (Lehmberg 2008; Pasternak 2009). Finally, nitrous oxide is a proven risk factor for nausea and vomiting (Apfel 2004).

Description of the intervention

As a weak anaesthetic, nitrous oxide is generally not used alone in general anaesthesia. Although there is considerable variation in how this drug is used, a typical scenario would be the maintenance of surgical anaesthesia, for whatever period required, by the administration of 69% nitrous oxide, 29% oxygen, and 2% of a potent volatile anaesthetic agent such as sevoflurane. Alternatively, an intravenous drug could be continuously infused while the patient breathes 70% nitrous oxide and 30% oxygen. The effect of nitrous oxide is to reduce the dose of either a volatile or intravenous anaesthetic that is required to maintain an appropriate level of anaesthesia.

How the intervention might work

As is the case with other gaseous anaesthetic agents, the exact mechanism of action of nitrous oxide is not completely understood. Theories include antagonism at both the N‐methyl‐D‐aspartate (NMDA) excitatory receptors and central nicotinic receptors; and a similar inhibitory effect at the two‐pore K⁺ channel TWIK‐related potassium channel‐1 (TREK‐1), a potassium channel involved in polymodal pain perception, to display analgesic, anxiolytic, and amnesic properties (Gruss 2004; Jevtović‐Todorović 1998; Yamakura 2000).

As suggested above, nitrous oxide is often used as one component of a balanced anaesthetic approach. This has several potential advantages including a reduction in the requirements for other agents, and consequently a reduced incidence and severity of any adverse effects of those agents, a rapid onset of anaesthetic effect, and a more rapid recovery of consciousness once the anaesthesia is discontinued (Becker 2008). These advantages need to be balanced against the potential disadvantages of nitrous oxide. Mechanistically, many of the adverse effects of nitrous oxide are ascribed to the inactivation of the cobalamin form of vitamin B12, by oxidation, thereby inhibiting the action of methionine synthase, folate metabolism, and deoxyribonucleic acid synthesis. All of these are important for protein production and DNA synthesis (Guirguis 1990; Perry 1983; Rowland 1995). Moreover, nitrous oxide depresses some white cells' ability to respond to various stimuli and reduces the growth of other white cell elements (mononucleocytes) (Kripke 1987).

Why it is important to do this review

As nitrous oxide administration brings both advantages and disadvantages, a systematic review will assist the individual anaesthetist in making the most appropriate choice of anaesthetic technique on an individual patient basis. The balance of risk versus benefit is likely to depend on many factors. The aim of this Cochrane review was to quantitatively evaluate if nitrous oxide was responsible for clinically significant adverse events following general anaesthesia that could be safely avoided by the use of alternative agents. This may have a wide impact on the conduct of general anaesthesia.

Objectives

To determine if nitrous oxide‐based anaesthesia results in similar outcomes to nitrous oxide‐free anaesthesia in adults undergoing surgery.

Methods

Criteria for considering studies for this review

Types of studies

We included only randomized controlled trials (RCTs).

Types of participants

We included participants, aged 18 years and older, undergoing surgery with standard general anaesthesia.

Types of interventions

Intervention:

General anaesthesia where nitrous oxide was part of the anaesthetic technique used for the induction or maintenance of general anaesthesia, or both.

Control:

General anaesthesia using a volatile anaesthetic or propofol‐based maintenance of anaesthesia but no nitrous oxide.

Types of outcome measures

Primary outcomes

Inhospital case fatality rate (number or proportion of deceased participants after a defined period following anaesthesia).

Secondary outcomes

1. Pulmonary complications:  1.1 Pneumonia: We accepted any definition used by the authors of included papers;  1.2 Pulmonary atelectasis: We accepted any definition used by the authors of included papers.

2. Heart complications:  2.1 Myocardial infarction: We accepted any definition of myocardial infarction used by the authors of included papers.

3. Neurological complications:  3.1 Stroke: We accepted any definition of stroke used by the authors of included papers. Where there was no definition, we accepted in the outcome any participant with new neurological signs (paralysis, weakness or speech difficulties) that persisted for 24 hours or leaded to early death.

4. Other complications:  4.1 Severe nausea and vomiting: We accepted any definition of severe nausea and vomiting made by the authors of included trials. Where there was no definition, we accepted into the outcome any participant with at least two episodes of vomiting or who required at least three doses of antiemetic medication within 24 hours of surgery;  4.2 Venous thromboembolism: We accepted any definition of deep venous thrombosis or pulmonary embolism used by the authors of included papers;  4.3 Wound infection rate: We accepted any definition of wound infection made by the authors of included trials.

5. Length of stay:  5.1 Length of hospital stay: We accepted any definition of length of hospital stay made by the authors of included trials;  5.2 Length of intensive care unit (ICU) stay: We accepted any definition of length of ICU stay made by the authors of included trials.

Search methods for identification of studies

Electronic searches

We developed a specific strategy for each database (Appendix 1 for CENTRAL; Appendix 2 for MEDLINE; Appendix 3 for EMBASE; and Appendix 4 for ISI Web of Science).

Searching other resources

Two review authors (RS, WQJ) examined the reference lists of any retrieved articles for additional relevant publications. In addition, two review authors (BM, YL) manually searched conference proceedings and review articles for relevant studies. We contacted relevant trial authors to identify any additional or ongoing studies. We also searched for relevant trials on specific websites: http://clinicaltrials.gov/; http://controlled‐trials.com/; and http://www.centerwatch.com. We did not apply any language restrictions.

Data collection and analysis

Two review authors (RS, WQJ) developed and used a standardized data extraction form in accordance with the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Two review authors (RS, XFL) independently checked and entered data into RevMan 5.3 for statistical analysis.

Selection of studies

One review author (WQJ) scanned the titles and abstracts of articles retrieved by the search and removed those that did not meet our inclusion criteria. Three review authors (JHT, WQJ, RS) retrieved the full text of all potentially eligible studies. Two review authors (RS, WQJ) independently examined the full text articles for compliance with the inclusion criteria and selected studies eligible for inclusion in the review. We resolved any disagreement as to study eligibility by discussion with a third review author (KHY).

Data extraction and management

We extracted data from eligible studies using a data form we had designed and pilot‐tested (Appendix 5). When a study either overlapped or was a duplicate of another study, WQ Jia and P Zhang contacted the study authors for clarification and, if confirmed, used the publication with the more detailed data for this systematic review and combined the additional data. Two review authors (RS, PZ) contacted the original study authors for additional data for included outcomes that were not published in the study. Two review authors (WQJ, RS) independently extracted the data and resolved any disagreement by consulting a third review author (KHY).

We extracted the following information:

Study design (RCT).
Participants (number, age, gender, American Society of Anesthesiologists (ASA) physical status classification, disease, type of surgery).
Intervention (concentration of nitrous oxide, mixed inhaled anaesthetic, concentration of oxygen, duration of inhaled nitrous oxide).
Quality assessment (sequence generation, allocation concealment, blinding, incomplete outcome data, other issues).
Outcome (primary and secondary outcomes, methods used to assess outcomes, time of follow‐up).

Assessment of risk of bias in included studies

Two review authors (RS, BM) independently assessed the quality of the studies by constructing a 'Risk of bias' table for each study which included sequence generation, allocation concealment, blinding, incomplete outcome data, selective reporting, and other bias (Higgins 2011). Any disagreements were resolved by discussion between the two review authors.

We assessed the quality factors of each study separately. These were classified as either 'low', 'high', or 'unclear' risk of bias.

Measures of treatment effect

Considering dichotomous variables, we expressed the difference in the number of events in the nitrous oxide‐based group and the nitrous oxide‐free group as an odds ratio (OR) for complications and Peto odds ratio (Peto OR) for the inhospital case fatality rate. For length of stay, we only pooled the data expressed as mean and standard deviation (SD). The effect size for length of stay was the mean difference (MD). We presented 95% confidence intervals (CIs) for all outcomes.

Unit of analysis issues

Non‐standard design RCTs can present statistical problems. Whilst we did not anticipate including crossover or cluster randomized designs in this Cochrane review, we expected multiple intervention groups. We took care to avoid 'unit of analysis' errors when analysing these types of trials (Higgins 2011).

Dealing with missing data

In the event of missing data, two review authors (WQJ, RS) tried to contact the authors of the original studies in order to obtain the necessary information. Two review authors (XFL, RS) analysed the data on an intention‐to‐treat (ITT) basis as far as possible.

Assessment of heterogeneity

We considered whether the clinical and methodological characteristics of the included studies were sufficiently similar for meta‐analysis to provide a meaningful summary. Statistically, we examined heterogeneity with the Chi² test and by calculating the I² statistic. We considered heterogeneity to be substantial when the I² statistic > 50% and carefully considered the data before reporting any pooled results (Higgins 2002). If substantial heterogeneity was detected, we explored possible explanations in subgroup analyses.

Assessment of reporting biases

We conducted a comprehensive search for eligible studies. If there were 10 or more studies in an analysis, we used a funnel plot to explore the possibility of publication bias and other reporting biases. In the analyses for dichotomous outcomes we also assessed publication bias statistically with the use of Egger's test (Egger 1997) performed with Stata 11.0. We based evidence of asymmetry on P < 0.05.

Data synthesis

We used meta‐analysis for data synthesis. We used a fixed‐effect model if the measure of inconsistency was low for all comparisons (I² statistic < 50%); otherwise we used a random‐effects model for measures with high inconsistency. Where we did not conduct meta‐analysis, we described the findings of the included studies qualitatively.

We included the following outcomes in the 'Summary of findings' tables:

Inhospital case fatality rate.
Pneumonia.
Pulmonary atelectasis.
Myocardial infarction.
Stroke.
Severe nausea and vomiting.
Length of hospital stay.
Venous thromboembolism.
Wound infection rate.

We rated the quality of evidence for each outcome following the guidelines of Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach (Schünemann 2009) and based on the following five downgrade factors: risk of bias, inconsistency, indirectness, imprecision, and publication bias. For each downgrade factor, a judgment of 'no', 'serious (downgrade the quality of evidence by one level)', or 'very serious (downgrade the quality of evidence by two levels)' was assigned. At the very beginning, we classified all the outcomes as at 'high' quality by default, and after rating, each outcome could receive a grade of either 'high', 'moderate', 'low', or 'very low' quality.

Subgroup analysis and investigation of heterogeneity

We undertook subgroup analyses as follows, as stated in the Cochrane protocol (Yang 2011):

Type of surgery (day‐case procedures/examinations versus intra‐abdominal surgery versus neurosurgery versus vascular surgery versus ophthalmic surgery versus breast surgery).
Different concentrations of inhaled nitrous oxide (high concentration [higher than 50%] versus low concentration [equal to or lower than 50%]).
Different intervention in the nitrous oxide‐free group (propofol‐based maintenance of anaesthesia versus volatile anaesthetic‐based maintenance of anaesthesia).

Sensitivity analysis

To evaluate whether the results of the systematic review were robust, we conducted sensitivity analyses based on the methodological quality (high quality versus low quality) and the percentages of withdrawals (above 10% versus below 10%) of the included RCTs.

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies; and Characteristics of studies awaiting classification sections.

Results of the search

The number of potential RCTs screened for inclusion in this Cochrane review is outlined in the study flow diagram (Figure 1).

We identified a total of 2906 references through searches of electronic databases and a further 229 through other resources. After removing the duplicates, we screened 2127 unique references. We excluded 2040 records based on titles and abstracts, and a further eight studies are awaiting classification (see Characteristics of studies awaiting classification) as we were unable to obtain their full texts from either our university library, the Danish National Library, or Cochrane Anaesthesia, Critical and Emergency Care Group members. We assessed 79 full text papers, of which 38 reports (consisting of 35 trials) were eligible for inclusion in this Cochrane review.

Included studies

We included 35 trials in this Cochrane review; see Characteristics of included studies.

Four studies included participants who had undergone day‐case procedures or examinations (Arellano 2000; Sengupta 1988; Short 1985; Van Hemelrijck 1991); 14 studies included participants who had undergone intra‐abdominal surgery (Akca 2004; Brodsky 2005; Chen 2013; Fleischmann 2005; Jensen 1992; Jensen 1993a; Jensen 1993b; Krogh 1994; Lee 2005; Lonie 1986; Mraovic 2008; Paredi 1994; Pedersen 1993; Sukhani 1994); three studies included participants who had undergone neurosurgery (Lampe 1990; Singh 2011; Todd 1993); two studies included participants who had undergone vascular surgery (Badner 2000; Kozmary 1990); one study included participants who had undergone ophthalmic surgery (Deleu 2000); one study included participants who had undergone breast surgery (Vanacker 1999); one study included participants who had undergone orthopedic surgery (Alhashemi 1997); and one study included participants who had undergone thoracic surgery (Yoshimura 2014). Eight studies included participants who had undergone different types of surgery (Bloomfield 1988; Eger 1990; ENIGMA II trial 2014; ENIGMA trial 2007; Gilani 2008; Larsen 2000; Leung 2006; Myles 2008a).

Twenty‐six studies used high concentrations of nitrous oxide in the nitrous oxide‐based group (Akca 2004; Alhashemi 1997; Arellano 2000; Badner 2000; Chen 2013; Eger 1990; ENIGMA II trial 2014; ENIGMA trial 2007; Fleischmann 2005; Gilani 2008; Jensen 1992; Jensen 1993a; Jensen 1993b; Kozmary 1990; Krogh 1994; Lampe 1990; Larsen 2000; Lee 2005; Lonie 1986; Myles 2008a; Pedersen 1993; Short 1985; Singh 2011; Todd 1993; Van Hemelrijck 1991; Sukhani 1994); three studies used low concentrations of nitrous oxide (Brodsky 2005; Mraovic 2008; Yoshimura 2014); and one study used both low and high concentrations of nitrous oxide (Sengupta 1988). Five studies did not report the concentration of nitrous oxide (Bloomfield 1988; Deleu 2000; Leung 2006; Paredi 1994; Vanacker 1999).

Ten studies used propofol‐based maintenance of anaesthesia in the nitrous oxide‐free group (Alhashemi 1997; Arellano 2000; Deleu 2000; Jensen 1992; Jensen 1993b; Larsen 2000; Krogh 1994; Sukhani 1994; Todd 1993; Yoshimura 2014); 22 studies used volatile anaesthetic‐based maintenance of anaesthesia in the nitrous oxide‐free group (Akca 2004; Badner 2000; Bloomfield 1988; Brodsky 2005; Chen 2013; Eger 1990; Fleischmann 2005; Gilani 2008; Jensen 1993a; Kozmary 1990; Lampe 1990; Lee 2005; Leung 2006; Lonie 1986; Mraovic 2008; Paredi 1994; Pedersen 1993; Sengupta 1988; Short 1985; Singh 2011; Vanacker 1999; Van Hemelrijck 1991). Three studies used different techniques of anaesthesia in the nitrous oxide‐free group (ENIGMA II trial 2014; ENIGMA trial 2007; Myles 2008a).

Of the 35 included trials, 24 trials reported outcomes identified as of interest for this review (Arellano 2000; Chen 2013; Deleu 2000; Eger 1990; ENIGMA II trial 2014; ENIGMA trial 2007; Fleischmann 2005; Jensen 1992; Jensen 1993a; Jensen 1993b; Kozmary 1990; Krogh 1994; Lampe 1990; Leung 2006; Mraovic 2008; Myles 2008a; Paredi 1994; Pedersen 1993; Sengupta 1988; Short 1985; Singh 2011; Todd 1993; Vanacker 1999; Van Hemelrijck 1991). Of the 11 trials excluded from the quantitative analysis, three reported quality of recovery (Brodsky 2005; Larsen 2000; Sukhani 1994); two reported non‐severe nausea and vomiting (Bloomfield 1988; Lonie 1986); one reported myocardial ischaemia (Badner 2000); one reported bowel distension (Akca 2004); one reported costs of anaesthesia and postoperative care (Alhashemi 1997); one reported postoperative pain (Gilani 2008); one reported postoperative opioid consumption (Lee 2005); and one reported lung collapse score (Yoshimura 2014).

Excluded studies

We excluded 41 studies after full text assessment. We excluded six of those studies because they were not RCTs (Antonini 1994; Barr 1999; Divatia 1996; Dover 1994; Morimoto 1997; Wesner 2005); 11 for including participants aged lower than 18 years (Jastak 1973; Johnson 1997; Lim 1992; Losasso 1992; Nightingale 1992; Ogg 1983; Rocca 2000; Saïssy 2000; Taki 2003; Towey 1979; Van den Berg 1995); six for including participants not undergoing general anaesthesia (Atanassoff 1994; Castéra 2001; Haraguchi 1995; Heath 1996; Kryshtalskyj 1990; Masood 2002); and 18 for using nitrous oxide in the control group (Atassi 2005; Bronco 2010; Cheong 2000; Einarsson 1997; Fredman 1998; Gozdemir 2007; Haessler 1993; Holst 1993; Ishii 1994; Jellish 1996; Nishiyama 1998; Simpson 1977; Sinha 2006; Smith 1993; Vari 2010; Yamakage 2001; Yang 2004; Zuurmond 1986). See Characteristics of excluded studies.

Studies awaiting classification

Eight studies are awaiting classification (Adams 1994; Miralles Pardo 1991; Moussa 1995; Rashchupkin 2011; Röpcke 2001; Schaffranietz 2000; Segatto 1993; Shulunov 2002). We were unable to obtain full text articles of these eight publications from our university library, the Danish National Library, and Cochrane Anaesthesia, Critical and Emergency Care Group members. Of these eight studies, seven were published in non‐English languages (three studies were in German, two studies were in Russian, one study was in Italian, and one study was in Spanish). See Characteristics of studies awaiting classification.

Ongoing studies

We did not identify any ongoing studies.

Risk of bias in included studies

We have summarized our 'Risk of bias' assessments for each included study in Figure 2 and as percentages across all studies in Figure 3. The details and reasons for each assessment are listed in the Characteristics of included studies section. Seven studies were at low risk of bias (Akca 2004; Arellano 2000; ENIGMA II trial 2014; ENIGMA trial 2007; Fleischmann 2005; Lee 2005; Leung 2006).

'Risk of bias' summary: review authors' judgements about each 'Risk of bias' item for each included study.

'Risk of bias' graph: review authors' judgements about each 'Risk of bias' item presented as percentages across all included studies.

Allocation

All included studies mentioned randomization in the methodology, but only 16 trials stated the actual method used for randomization (Akca 2004; Alhashemi 1997; Arellano 2000; Badner 2000; Bloomfield 1988; Brodsky 2005; Chen 2013; ENIGMA II trial 2014; ENIGMA trial 2007; Fleischmann 2005; Lee 2005; Leung 2006; Mraovic 2008; Myles 2008a; Singh 2011; Yoshimura 2014).

In 23 studies, the trial authors did not give the details of the method of concealment of allocation, and we categorized these studies as 'unclear'. Concealment was adequate in 12 studies (Akca 2004; Arellano 2000; Chen 2013; ENIGMA II trial 2014; ENIGMA trial 2007; Fleischmann 2005; Jensen 1992; Lee 2005; Leung 2006; Myles 2008a; Todd 1993; Vanacker 1999).

Blinding

Participants and personnel were blinded in eight studies (Akca 2004; Arellano 2000; ENIGMA II trial 2014; ENIGMA trial 2007; Fleischmann 2005; Myles 2008a; Pedersen 1993; Singh 2011); four studies were not blinded (Larsen 2000; Paredi 1994; Sukhani 1994; Van Hemelrijck 1991); and the remaining studies were unclear.

We have separated 'blinding of outcome assessment (detection bias)' by type of outcome as the impact of outcome assessor knowledge of allocation may vary across different outcomes.

We assessed the 13 studies reporting clinical endpoints of inhospital case fatality rate or length of stay as being at a low risk of detection bias, since the outcome measurements were unlikely to have been influenced by lack of blinding (Chen 2013; Eger 1990; ENIGMA II trial 2014; ENIGMA trial 2007; Fleischmann 2005; Jensen 1992; Jensen 1993a; Krogh 1994; Lampe 1990; Leung 2006; Myles 2008a; Singh 2011; Todd 1993). Of the 32 studies reporting clinical endpoints of complications, the outcome assessors were blinded in 25 studies (Akca 2004; Alhashemi 1997; Arellano 2000; Badner 2000; Bloomfield 1988; Brodsky 2005; Chen 2013; Eger 1990; ENIGMA II trial 2014; ENIGMA trial 2007; Fleischmann 2005; Jensen 1993a; Jensen 1993b; Kozmary 1990; Lampe 1990; Larsen 2000; Lee 2005; Lonie 1986; Mraovic 2008; Paredi 1994; Pedersen 1993; Singh 2011; Sukhani 1994; Van Hemelrijck 1991; Yoshimura 2014); outcome assessors were not blinded in one study (Short 1985); and the remaining studies were unclear (Deleu 2000; Gilani 2008; Jensen 1992; Sengupta 1988; Todd 1993; Vanacker 1999).

Incomplete outcome data

The number of participants entering the trials and the number subjected to analysis, as mentioned in the results, were the same in 21 studies (Akca 2004; Alhashemi 1997; Bloomfield 1988; Brodsky 2005; Gilani 2008; Jensen 1992; Jensen 1993b; Kozmary 1990; Krogh 1994; Lampe 1990; Larsen 2000; Lee 2005; Leung 2006; Lonie 1986; Myles 2008a; Paredi 1994; Sukhani 1994; Todd 1993; Vanacker 1999; Van Hemelrijck 1991; Yoshimura 2014). Of the 14 studies that had withdrawals, the missing outcome data was balanced in numbers across the intervention groups. Nine trials gave similar reasons for missing data across groups (Arellano 2000; Badner 2000; Chen 2013; ENIGMA II trial 2014; ENIGMA trial 2007; Fleischmann 2005; Jensen 1993a; Mraovic 2008; Singh 2011). The remaining five studies had insufficient information to enable us to form a judgment (Deleu 2000; Eger 1990; Pedersen 1993; Sengupta 1988; Short 1985).

Selective reporting

Two studies, ENIGMA II trial 2014 and ENIGMA trial 2007, were registered on ClinicalTrials.gov (NCT00430989 and NCT00164047, respectively). The study protocols were available and all of the pre‐specified (primary and secondary) outcomes that were of interest in the review were reported in the pre‐specified way. Of the 33 studies that had no protocol, one study had not reported all the pre‐specified primary outcomes (Myles 2008a); and the remaining studies reported all the outcomes described in their method sections.

Other potential sources of bias

Given the outcomes of interest in this Cochrane review, such as inhospital death and complications, were at low incidence, most of the included trials were underpowered for these outcomes. We assessed this item as high risk in studies that reported the outcomes of inhospital death or complications, but had fewer than 50 participants per arm (Chaparro 2013). Therefore we assessed 13 studies as at high risk of bias (Chen 2013; Deleu 2000; Jensen 1992; Jensen 1993a; Jensen 1993b; Kozmary 1990; Lampe 1990; Pedersen 1993; Sengupta 1988; Short 1985; Singh 2011; Vanacker 1999; Van Hemelrijck 1991).

Effects of interventions

See: Table 1; Table 2

Summary of findings for the main comparison. Summary of findings’ table 1.

Nitrous oxide‐based compared to nitrous oxide‐free for general anaesthesia
Patient or population: adult patients 18 years and above undergoing standard general anaesthesia  Settings: operating room  Intervention: nitrous oxide‐based techniques  Comparison: nitrous oxide‐free techniques
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	nitrous oxide‐free	Nitrous oxide‐based
Inhospital case fatality rate	Study population		OR 0.87   (0.61 to 1.26)	10148  (8 studies)	⊕⊕⊕⊝  moderate¹	—
	12 per 1000	11 per 1000   (8 to 16)
	Moderate
	0 per 1000	0 per 1000   (0 to 0)
Pneumonia	Study population		OR 1.68   (1 to 2.81)	2699  (8 studies)	⊕⊕⊝⊝  low^2,3	The sensitivity analysis suggested that the results of meta‐analysis was not robust.
	17 per 1000	27 per 1000   (17 to 45)
	Moderate
	11 per 1000	18 per 1000   (11 to 30)
Pulmonary atelectasis	Study population		OR 1.57   (1.18 to 2.1)	2400  (5 studies)	⊕⊕⊕⊕  high	—
	79 per 1000	119 per 1000   (92 to 153)
	Moderate
	50 per 1000	76 per 1000   (58 to 100)
Myocardial infarction	Study population		OR 1.01   (0.84 to 1.22)	9246  (6 studies)	⊕⊕⊕⊕  high	—
	51 per 1000	51 per 1000   (43 to 61)
	Moderate
	65 per 1000	66 per 1000   (55 to 78)
Stroke	Study population		OR 1.47   (0.86 to 2.53)	9142  (4 studies)	⊕⊕⊕⊝  moderate³	—
	5 per 1000	7 per 1000   (4 to 12)
	Moderate
	3 per 1000	4 per 1000   (3 to 8)
Severe nausea and vomiting	Study population		OR 1.44   (0.97 to 2.15)	11045  (10 studies)	⊕⊕⊝⊝  low^4,5	The sensitivity analysis suggested that the results of meta‐analysis was not robust.
	95 per 1000	131 per 1000   (92 to 184)
	Moderate
	108 per 1000	148 per 1000   (105 to 207)
Length of hospital stay		The mean length of hospital stay in the intervention groups was  0.36 higher   (0.69 lower to 1.4 higher)		1103  (6 studies)	⊕⊕⊕⊝  moderate⁵	—
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; OR: odds ratio.
GRADE Working Group grades of evidence  High quality: further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: we are very uncertain about the estimate.

First author	Journal/Conference proceedings etc	Year

RCT	Relevant participants (age over 18 years, having standard general anaesthesia)	Relevant interventions (nitrous oxide‐based throughout duration of anaesthesia compared with nitrous oxide‐free)	Relevant outcomes:  Inhospital case fatality rate, pulmonary complications (pneumonia and pulmonary atelectasis), heart complications (myocardial infarction), neurological complications (stroke), other complications (venous thromboembolism, wound infection rate and severe nausea and vomiting), length of stay (length of hospital stay and length of ICU stay)
Yes/No/Unclear	Yes/No/Unclear	Yes/No/Unclear	Yes/No/Unclear

Code each paper	Author(s)	Journal/Conference proceedings etc	Year
A	The paper listed above
B	Further papers

Participant and trial characteristics
Characteristics	Further details
Age (mean, median, range, etc)
Number of participants in each intervention group
Sex of participants (numbers/%, etc)
Disease status/type, etc? (if applicable)
ASA physical status classification
Type of surgery
Single centre/multi‐centre
Country/Countries
Number of participants who were analysed
Trial design (e.g. parallel/cross‐over*)
Other

Groups	Details of intervention
Nitrous oxide based group(concentration of inhaled N₂O, O₂ separately, duration of inhaled N₂O)
Nitrous oxide‐free group (inhaled gas, concentration of inhaled O₂)

Allocation of intervention
State here method used to generate allocation and reasons for grading	Grade (circle)
Comment on allocation by review authors or included study quote concerning allocation	Adequate (random)
	Inadequate (e.g. alternate)
	Unclear
Concealment of allocation
State here method used to generate allocation and reasons for grading	Grade (circle)
Comment on allocation by review authors or included study quote concerning allocation	Adequate (random)
	Inadequate (e.g. alternate)
	Unclear
Blinding
Person responsible for participants' care	Yes/No
Participant	Yes/No
Outcome assessor	Yes/No
Other (please specify)	Yes/No
Intention‐to‐treat
All participants entering trial
15% or fewer excluded
More than 15% excluded
Not analysed as "intention‐to‐treat"
Unclear

For "duration of stay" data
Code of paper	Outcomes	Unit of measurement	N₂O‐based group		N₂O‐free group		Details if outcome only described in text
Code of paper	Outcomes	Unit of measurement	n	Mean	n	Mean	Details if outcome only described in text
Secondary outcomes
	Duration of hospital stay						Yes/No
	Duration of ICU stay

For dichotomous data
Code of paper	Outcomes	N₂O‐based group(n)  n = number of participants, not number of events	N₂O‐free group(n)  n = number of participants, not number of events
Primary outcomes
	Inhospital case fatality rate
Secondary outcomes
	Pneumonia
	Pulmonary atelectasis
	Myocardial infarction
	Stroke
	Severe nausea and vomiting
	Venous thromboembolism
	Wound infection rate

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Inhospital case fatality rate	8	10148	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.87 [0.61, 1.26]
2 Pneumonia	8	2699	Odds Ratio (M‐H, Fixed, 95% CI)	1.68 [1.00, 2.81]
3 Pulmonary atelectasis	5	2400	Odds Ratio (M‐H, Fixed, 95% CI)	1.57 [1.18, 2.10]
4 Myocardial infarction	6	9246	Odds Ratio (M‐H, Fixed, 95% CI)	1.01 [0.84, 1.22]
5 Stroke	4	9142	Odds Ratio (M‐H, Fixed, 95% CI)	1.47 [0.86, 2.53]
6 Severe nausea and vomiting	10	11045	Odds Ratio (M‐H, Random, 95% CI)	1.44 [0.97, 2.15]
7 Venous thromboembolism	2	9004	Odds Ratio (M‐H, Fixed, 95% CI)	0.73 [0.45, 1.20]
8 Wound infection rate	6	9789	Odds Ratio (M‐H, Random, 95% CI)	1.22 [0.84, 1.78]
9 Length of hospital stay	6	1103	Mean Difference (IV, Random, 95% CI)	0.36 [‐0.69, 1.40]
10 Inhospital case fatality rate: type of surgery	4	646	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.64 [0.11, 3.88]
10.1 Intra‐abdominal surgery	2	499	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.41 [0.06, 3.03]
10.2 Neurosurgery	2	147	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.45 [0.07, 287.21]
11 Pneumonia: type of surgery	6	427	Odds Ratio (M‐H, Fixed, 95% CI)	1.29 [0.48, 3.48]
11.1 Intra‐abdominal surgery	3	193	Odds Ratio (M‐H, Fixed, 95% CI)	1.12 [0.33, 3.78]
11.2 Neurosurgery	3	234	Odds Ratio (M‐H, Fixed, 95% CI)	1.71 [0.30, 9.81]
12 Pulmonary atelectasis: type of surgery	3	128	Odds Ratio (M‐H, Fixed, 95% CI)	0.16 [0.02, 1.06]
12.1 Intra‐abdominal surgery	2	102	Odds Ratio (M‐H, Fixed, 95% CI)	0.16 [0.02, 1.06]
12.2 Neurosurgery	1	26	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
13 Myocardial infarction: type of surgery	3	212	Odds Ratio (M‐H, Fixed, 95% CI)	1.24 [0.42, 3.67]
13.1 Intra‐abdominal surgery	1	91	Odds Ratio (M‐H, Fixed, 95% CI)	2.07 [0.48, 8.93]
13.2 Neurosurgery	1	87	Odds Ratio (M‐H, Fixed, 95% CI)	3.44 [0.14, 86.92]
13.3 Vascular surgery	1	34	Odds Ratio (M‐H, Fixed, 95% CI)	0.25 [0.02, 2.74]
14 Stroke: type of surgery	2	138	Odds Ratio (M‐H, Fixed, 95% CI)	2.46 [0.53, 11.48]
14.1 Neurosurgery	1	87	Odds Ratio (M‐H, Fixed, 95% CI)	1.74 [0.28, 10.95]
14.2 Ophthalmic surgery	1	51	Odds Ratio (M‐H, Fixed, 95% CI)	5.20 [0.24, 113.98]
15 Severe nausea and vomiting: type of surgery	8	2041	Odds Ratio (M‐H, Fixed, 95% CI)	1.13 [0.72, 1.78]
15.1 Day‐case procedure/examination	4	1624	Odds Ratio (M‐H, Fixed, 95% CI)	0.65 [0.30, 1.41]
15.2 Intra‐abdominal surgery	3	233	Odds Ratio (M‐H, Fixed, 95% CI)	1.56 [0.65, 3.71]
15.3 Breast surgery	1	184	Odds Ratio (M‐H, Fixed, 95% CI)	1.44 [0.70, 2.99]
16 Wound infection rate: type of surgery	3	525	Odds Ratio (M‐H, Random, 95% CI)	1.63 [0.28, 9.33]
16.1 Intra‐abdominal surgery	2	499	Odds Ratio (M‐H, Random, 95% CI)	1.63 [0.28, 9.33]
16.2 Neurosurgery	1	26	Odds Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
17 Length of hospital stay: type of surgery	3	556	Mean Difference (IV, Fixed, 95% CI)	‐0.45 [‐1.44, 0.54]
17.1 Intra‐abdominal surgery	2	530	Mean Difference (IV, Fixed, 95% CI)	‐0.12 [‐1.25, 1.00]
17.2 Neurosurgery	1	26	Mean Difference (IV, Fixed, 95% CI)	‐1.60 [‐3.71, 0.51]
18 Inhospital case fatality rate: concentrations of inhaled nitrous oxide	7	9920	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.86 [0.60, 1.24]
18.1 High‐concentration nitrous oxide	7	9920	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.86 [0.60, 1.24]
19 Stroke: concentrations of inhaled nitrous oxide	3	9091	Odds Ratio (M‐H, Fixed, 95% CI)	1.39 [0.80, 2.42]
19.1 High‐concentration nitrous oxide	3	9091	Odds Ratio (M‐H, Fixed, 95% CI)	1.39 [0.80, 2.42]
20 Severe nausea and vomiting: concentrations of inhaled nitrous oxide	8	10847	Odds Ratio (M‐H, Random, 95% CI)	1.37 [0.89, 2.11]
20.1 High‐concentration nitrous oxide	7	10691	Odds Ratio (M‐H, Random, 95% CI)	1.34 [0.81, 2.19]
20.2 Low‐concentration nitrous oxide	2	156	Odds Ratio (M‐H, Random, 95% CI)	1.39 [0.53, 3.68]
21 Length of hospital stay: concentrations of inhaled nitrous oxide	5	875	Mean Difference (IV, Random, 95% CI)	0.45 [‐1.03, 1.93]
21.1 High‐concentration nitrous oxide	5	875	Mean Difference (IV, Random, 95% CI)	0.45 [‐1.03, 1.93]
22 Inhospital case fatality rate: techniques used in the nitrous oxide‐free group	6	1144	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.33 [0.29, 6.00]
22.1 Propofol‐based maintenance of anaesthesia used in the nitrous oxide‐free group	1	121	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.45 [0.07, 287.21]
22.2 Volatile anaesthetic‐based maintenance of anaesthesia used in the nitrous oxide‐free group	5	1023	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.11 [0.22, 5.58]
23 Pneumonia: techniques used in the nitrous oxide‐free group	7	687	Odds Ratio (M‐H, Fixed, 95% CI)	1.13 [0.45, 2.86]
23.1 Propofol‐based maintenance of anaesthesia used in the nitrous oxide‐free group	2	181	Odds Ratio (M‐H, Fixed, 95% CI)	0.50 [0.07, 3.61]
23.2 Volatile anaesthetic‐based maintenance of anaesthesia used in the nitrous oxide‐free group	5	506	Odds Ratio (M‐H, Fixed, 95% CI)	1.42 [0.49, 4.06]
24 Pulmonary atelectasis: techniques used in the nitrous oxide‐free group	4	388	Odds Ratio (M‐H, Fixed, 95% CI)	0.32 [0.09, 1.12]
24.1 Propofol‐based maintenance of anaesthesia used in the nitrous oxide‐free group	2	102	Odds Ratio (M‐H, Fixed, 95% CI)	0.16 [0.02, 1.06]
24.2 Volatile anaesthetic‐based maintenance of anaesthesia used in the nitrous oxide‐free group	2	286	Odds Ratio (M‐H, Fixed, 95% CI)	0.70 [0.12, 4.29]
25 Myocardial infarction: techniques used in the nitrous oxide‐free group	4	242	Odds Ratio (M‐H, Fixed, 95% CI)	0.96 [0.37, 2.53]
25.1 Volatile anaesthetic‐based maintenance of anaesthesia used in the nitrous oxide‐free group	4	242	Odds Ratio (M‐H, Fixed, 95% CI)	0.96 [0.37, 2.53]
26 Stroke: techniques used in the nitrous oxide‐free group	2	138	Odds Ratio (M‐H, Fixed, 95% CI)	2.46 [0.53, 11.48]
26.1 Propofol‐based maintenance of anaesthesia used in the nitrous oxide‐free group	1	51	Odds Ratio (M‐H, Fixed, 95% CI)	5.20 [0.24, 113.98]
26.2 Volatile anaesthetic‐based maintenance of anaesthesia used in the nitrous oxide‐free group	1	87	Odds Ratio (M‐H, Fixed, 95% CI)	1.74 [0.28, 10.95]
27 Severe nausea and vomiting: techniques used in the nitrous oxide‐free group	8	2041	Odds Ratio (M‐H, Fixed, 95% CI)	1.13 [0.72, 1.78]
27.1 Propofol‐based maintenance of anaesthesia used in the nitrous oxide‐free group	1	1417	Odds Ratio (M‐H, Fixed, 95% CI)	1.99 [0.18, 22.04]
27.2 Volatile anaesthetic‐based maintenance of anaesthesia used in the nitrous oxide‐free group	7	624	Odds Ratio (M‐H, Fixed, 95% CI)	1.11 [0.70, 1.75]
28 Wound infection rate: techniques used in the nitrous oxide‐free group	4	785	Odds Ratio (M‐H, Random, 95% CI)	2.13 [0.44, 10.22]
28.1 Volatile anaesthetic‐based maintenance of anaesthesia used in the nitrous oxide‐free group	4	785	Odds Ratio (M‐H, Random, 95% CI)	2.13 [0.44, 10.22]
29 Length of hospital stay: techniques used in the nitrous oxide‐free group	5	1044	Mean Difference (IV, Fixed, 95% CI)	0.20 [‐0.36, 0.75]
29.1 Volatile anaesthetic‐based maintenance of anaesthesia used in the nitrous oxide‐free group	5	1044	Mean Difference (IV, Fixed, 95% CI)	0.20 [‐0.36, 0.75]

Methods	Multi‐centre RCT
Participants	Setting: two Viennese hospitals Inclusion criteria: ASA I‐III patients, 18 to 80 years of age, scheduled for elective colon resection scheduled to last more than 2 hours Exclusion criteria: patients with bowel obstruction or having minor colon surgery (e.g. polypectomy, isolated colostomy) Participant numbers: 344 randomly assigned; 344 analysed
Interventions	Intervention: anaesthetic management was standardized. Sodium thiopental (3 to 5 mg/kg) and vecuronium (0.1 mg/kg) were used for induction; anaesthesia subsequently was maintained with isoflurane (0.5 to 1.0% in 65% nitrous oxide), vecuronium, and remifentanil (0.2 mg/kg/min).  Control: anaesthetic management was standardized. Sodium thiopental (3 to 5 mg/kg) and vecuronium (0.1 mg/kg) were used for induction; anaesthesia subsequently was maintained with isoflurane (0.5 to 1.0% in air), vecuronium, and remifentanil (0.2 mg/kg/min).
Outcomes	Other outcomes:   Bowel distension
Notes	—
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were assigned to one of two groups using a reproducible set of computer‐generated random numbers."
Allocation concealment (selection bias)	Low risk	Quote: "The assignments were kept in sealed, sequentially numbered opaque envelopes that were opened after induction of anaesthesia."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Great care was exercised to prevent the surgeons from observing the administered gas mixture."
Blinding of outcome assessment (detection bias)   Inhospital case fatality rate/length of stay	Unclear risk	—
Blinding of outcome assessment (detection bias)   Complications	Low risk	Quote: "The rater was blinded to anaesthesia management."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No missing data.
Selective reporting (reporting bias)	Low risk	All outcomes described in methods section reported.
Other bias	Low risk	No other potential sources of bias were detected.

Methods	Single‐centre RCT
Participants	Setting: the Ottawa General Hospital, Canada Inclusion criteria: ASA I‐II patients, scheduled to undergo arthroscopic knee surgery, and electing general anaesthesia Exclusion criteria: patient preference for regional anaesthesia; age < 20 or > 60 years; body mass index either < 20 or > 30 kg/m²; current or chronic use of benzodiazepines or other sedative‐hypnotics; excessive alcohol intake; moderate or severe cardiac or respiratory disease; severe or uncontrolled hypertension; known allergy to any of the study medications; or chronic use of drugs known to interfere with the metabolism or clinical effects of the study medications Participant numbers: 93 randomly assigned; 93 analysed
Interventions	Intervention: patients received nitrous oxide 70% supplemented with isoflurane 0.5 to 1.0% or with intermittent boluses of 7 to 15 μg/kg iv alfentanil every 10 to 15 min.  Control: patients received intermittent boluses of 7 to 15 μg/kg iv alfentanil every 10 to 15 min in conjunction with a continuous infusion of propofol.
Outcomes	Other outcomes:   Costs of anaesthesia and postoperative care
Notes	—
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were allocated according to a computer generated randomization schedule."
Allocation concealment (selection bias)	Unclear risk	No details given.
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	No details given.
Blinding of outcome assessment (detection bias)   Inhospital case fatality rate/length of stay	Unclear risk	—
Blinding of outcome assessment (detection bias)   Complications	Low risk	Quote: "Blinding was achieved by precluding the trained observer, who recorded all post‐operative data, from gaining any knowledge of the intra‐operative anaesthetic care."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No missing data.
Selective reporting (reporting bias)	Low risk	All outcomes described in methods section reported.
Other bias	Low risk	No other potential sources of bias were detected.

Methods	Multicentre RCT
Participants	Setting: 45 participating centres from 10 countries Inclusion criteria: Adult males and females age ≥ 45 years, undergoing noncardiac surgery and general anaesthesia expected to exceed two hours. At increased risk of cardiac events, defined as any of History of coronary artery disease as defined by a history of any one of the following: i. angina ii. MI iii. segmental wall motion abnormality on echocardiography or a fixed defect on radionuclide imaging iv. a positive exercise stress test for cardiac ischaemia v. a positive radionuclide exercise, echocardiographic exercise, or pharmacological cardiovascular stress test for cardiac ischaemia vi. coronary revascularization (CABG or PTCA) vii. angiographic evidence of atherosclerotic stenosis > 50% of the diameter of any coronary artery viii. ECG with pathological Q waves in two contiguous leads; Heart failure; Cerebrovascular disease thought due to atherothrombotic disease; Aortic or peripheral vascular disease; Or three or more of the following risk factors: Age ≥ 70 years; Any history of congestive heart failure; Diabetes and currently on an oral hypoglycaemic agent or insulin therapy; Current treatment for hypertension; Preoperative serum creatinine > 175 µmol/L (> 2.0 mg/dL); Current or previous high cholesterol ≥ 6.2 mmol/L (> 240 mg/dL); History of a TIA (i.e. a transient focal neurological deficit that lasted less than 24 hours and thought to be vascular in origin); Emergency/urgent surgery (i.e. surgery which must be undertaken within 24 hours of acute presentation to hospital); High‐risk type of surgery (i.e. intrathoracic or intraperitoneal). Exclusion criteria: Having cardiac surgery. Marked impairment of gas‐exchange expected to require FiO₂> 0.5 intraoperatively. Specific circumstances where nitrous oxide is contraindicated (e.g. volvulus, pulmonary hypertension, raised intracranial pressure) or the anaesthetist plans to use supplemental oxygen (e.g. colorectal surgery). Nitrous oxide unavailable for use. Participant numbers: 7112 randomly assigned; 6992 analysed
Interventions	Intervention: 70% nitrous oxide  Control: no nitrous oxide
Outcomes	Primary outcomes: The primary endpoint is a composite of death and cardiovascular events (clinical and silent MI, cardiac failure, cardiac arrest, pulmonary embolism, and stroke) measured at 30 days after surgery.    Secondary outcomes: Myocardial infarction: A typical rise of troponin or a typical fall of an elevated troponin with at least one value above the 99th percentile of the upper reference limit, detected at its peak‐post surgery in a patient without a documented alternative explanation for an elevated troponin (e.g. pulmonary embolism). This criterion also requires that 1 of the following must also exist: a) ischaemic signs or symptoms (i.e., chest, arm, neck, or jaw discomfort; shortness of breath, pulmonary edema); b) development of pathologic Q waves present in any two contiguous leads that are > 30 milliseconds; c) ECG changes indicative of ischaemia (i.e. ST segment elevation [> 2 mm in leads V1, V2, or V3; or > 1 mm in the other leads], ST segment depression [> 1 mm], or symmetric inversion of T waves > 1 mm) in at least two contiguous leads; d) coronary artery intervention (i.e. PCI or CABG surgery); e) new or presumed new cardiac wall motion abnormality on echocardiography or new or presumed new fixed defect on radionuclide imaging. Pathologic findings of an acute or healing myocardial infarction. Development of new pathological Q waves on an ECG if troponin levels were not obtained or were obtained at times that could have missed the clinical event. Wound infection At least one of the following: Purulent drainage from the incision. Positive microbial culture from the incision. Documentation of a wound infection in the medical record. Stroke New cerebral infarction or haemorrhage on CT scan, MRI or documented new neurological signs (paralysis, weakness or speech difficulties) lasting more than 24 hours or leading to earlier death (confirmed by a copy of the autopsy report or in the medical record). Severe nausea and vomiting At least two separate episodes of nausea or vomiting greater than six hours apart, or if requiring three or more doses of antiemetic medication, within three days after surgery. Pulmonary embolism A high probability VQ scan, spiral CT or documented on pulmonary angiogram; or pathological findings (with autopsy). Hospital stay
Notes	ClinicalTrials.gov identifier: NCT00430989
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was done with a computer‐generated code."
Allocation concealment (selection bias)	Low risk	Quote: "Randomization sequence was accessed via an automated telephone voice‐recognition service."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "The attending anaesthetists were aware of the patients' group assignments, but the patients, their surgical team were not."
Blinding of outcome assessment (detection bias)   Inhospital case fatality rate/length of stay	Low risk	Quote: "The postoperative interviewers, and endpoint adjudicators were unaware of the patients' group assignments."
Blinding of outcome assessment (detection bias)   Complications	Low risk	Quote: "The postoperative interviewers, and endpoint adjudicators were unaware of the patients' group assignments."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups.
Selective reporting (reporting bias)	Low risk	Protocol was registered as NCT00430989 and outcomes were reported.
Other bias	Low risk	No other potential sources of bias were detected.

Methods	RCT
Participants	Setting: unknown Inclusion criteria: 150 ASA I‐II patients, between 19 and 75 years old, undergoing elective laparoscopic gynaecological surgery (removal of ovarian tumours and cysts, myomectomy, laparoscopic‐assisted vaginal hysterectomy, and infertility surgery) Exclusion criteria: obesity (body mass index > 33 kg/m²), pregnancy, breast‐feeding, known hypersensitivity to drugs used in the study protocol, use of antiemetics, psychotropic drugs and steroids within 72 h before surgery; patients with known comorbidities that could increase the incidence of postoperative nausea and vomiting, i.e. diseases which impaired gastric motility (diabetes mellitus, chronic cholecystitis, gastric and intestinal disease, neuromuscular disorders, neuropathies, and liver dysfunction), vestibular disease, history of migraine headache, central nervous system injury, renal impairment, irregular menstrual cycle (duration of < 21 or > 35 days and/or variations between cycles > 4 days), alcoholism, and opioid addiction Participant numbers: 150 randomly assigned; 137 analysed
Interventions	Intervention: patients received 7.5 mg of midazolam by mouth 1 h before the surgery with no prophylactic antiemetics. After induction of anaesthesia with thiopental 5 mg/kg and fentanyl 1 to 2 μg/kg, patients were manually ventilated with oxygen via facemask. Endotracheal intubation was facilitated with vecuronium 0.1 mg/kg IV. Patients then received either 50% nitrous oxide with oxygen or 70% nitrous oxide with oxygen. Anaesthesia was maintained with sevoflurane (end‐tidal concentration approximately 1 MAC) and supplemental bolus doses of fentanyl intravenous (1 μg/kg) to keep heart rate and arterial blood pressure within 20% of baseline values and additional vecuronium was administered to maintain 1 or 2 twitches on the train‐of‐four monitor. All patients received 10 mL/kg of crystalloids intraoperatively.  Control: patients received 7.5 mg of midazolam by mouth 1 h before the surgery with no prophylactic antiemetics. After induction of anaesthesia with thiopental 5 mg/kg and fentanyl 1 to 2 μg/kg, patients were manually ventilated with oxygen via facemask. Endotracheal intubation was facilitated with vecuronium 0.1 mg/kg IV. Patients then received air and oxygen, FiO₂30%. Anaesthesia was maintained with sevoflurane (end‐tidal concentration approximately 1 MAC) and supplemental bolus doses of fentanyl intravenous (1 μg/kg) to keep heart rate and arterial blood pressure within 20% of baseline values and additional vecuronium was administered to maintain 1 or 2 twitches on the train‐of‐four monitor. All patients received 10 mL/kg of crystalloids intraoperatively.
Outcomes	Secondary outcomes: Severe nausea and vomiting: 2 or more episodes of vomiting and retching within a period of 30 min or total number of 3 or more emetic episodes during 24 h postoperatively.
Notes	13 patients were excluded from the analysis. 4 patients were excluded in nitrous oxide‐free group: 1 patient was treated with corticosteroids for urticaria at induction of anaesthesia, 1 patient had an anaesthesia time < 30 min, 2 patients had a protocol violation. 4 patients were excluded in nitrous oxide‐based group 1: 1 patient had a conversion to laparotomy, 1 patient's anaesthesia time was < 30 min, and 2 patients had a protocol violation. 5 patients were excluded from the nitrous oxide‐based group 2: 2 patients' surgery was converted to laparotomy, 1 patient each had severe hypotension after induction, which lasted more than 5 mins, acute coronary syndrome postoperatively, and anaesthesia time < 30 min.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomized by computer‐generated random numbers."
Allocation concealment (selection bias)	Unclear risk	No details given.
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	This was not reported.
Blinding of outcome assessment (detection bias)   Inhospital case fatality rate/length of stay	Unclear risk	—
Blinding of outcome assessment (detection bias)   Complications	Low risk	Quote: "Clinical nurses specifically trained for the study collected the data and were blinded to the anaesthesia technique used and randomizations."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups.
Selective reporting (reporting bias)	Low risk	All outcomes described in methods section reported.
Other bias	Low risk	No other potential sources of bias were detected.

Methods	Muli‐centre RCT
Participants	Setting: 2 hospitals from Australia and Hong Kong Inclusion criteria: patients undergoing elective noncardiac surgery, with risk factors or a known history of coronary artery disease (hypertension, diabetes, age older than 60 years, or preexisting history of coronary artery disease) Exclusion criteria: patients expected to require a high inspired oxygen concentration intraoperatively or with any relative contraindication to nitrous oxide (volvulus, pulmonary hypertension, increased intracranial pressure) Participant numbers: 59 randomly assigned; 59 analysed
Interventions	Intervention: patients had maintenance of general anaesthesia with nitrous oxide and FiO₂ 0.3, and one of three other hypnotic agents (isoflurane, sevoflurane, or propofol) at the discretion of the anaesthesiologist.   Control: patients had their anaesthesia maintained with FiO₂ 0.8 or FiO₂ 0.3, but without nitrous oxide.
Outcomes	Secondary outcomes: Length of hospital stay
Notes	—
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomly allocated via a computer‐generated random list."
Allocation concealment (selection bias)	Low risk	Quote: "The random sequence was concealed in opaque, sealed envelopes."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Administration and group identity were concealed from the surgeon and research staff."
Blinding of outcome assessment (detection bias)   Inhospital case fatality rate/length of stay	Low risk	The outcome measurement is not likely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Complications	Unclear risk	—
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No missing data.
Selective reporting (reporting bias)	High risk	Not all of the study's pre‐specified primary outcomes (e.g. myocardial infarction) had been reported.
Other bias	Low risk	No other potential sources of bias were detected.

Methods	RCT
Participants	Inclusion criteria: 60 female in‐patients (ASA I‐II), aged 18 to 65 years, scheduled for breast surgery with a duration of 1 to 3 hours Exclusion criteria: patients had body weight 20% outside normal weight, history of motion sickness or of postoperative nausea and vomiting, pregnant or breastfeeding patients, history of alcohol or drug abuse, sensitivity to narcotics, impaired renal or hepatic function, recent (< 30 days) participation in another study Participant numbers: 30 randomly assigned; 30 analysed
Interventions	Intervention: patients were breathing 100% oxygen with a fresh gas flow of 7 L/min for 2 to 3 minutes. A standardized anaesthetic technique consisting of propofol for induction (2 mg/kg) followed by desflurane with nitrous oxide for maintenance of anaesthesia was used in all patients. The concentration of anaesthetic given to the patients was based on previously determined MAC values and adjusted to the patient's needs as clinically indicated with the objective to maintain the heart rate and blood pressure within 20% of the baseline values. They received a pre‐induction dose of fentanyl 2 μg/kg; additional doses of fentanyl 1 μg/kg were given if there were signs of inadequate anaesthesia (i.e. movement, swallowing, tearing, salivation) despite changes in inhalation concentration. Muscle relaxation for intubation was achieved by a single dose of vecuronium 0.1 mg/kg. Mechanical ventilation was instituted in all patients and ventilatory settings were adjusted to achieve normocapnia; the fresh gas flow was reduced to 2 L/ min during maintenance of anaesthesia. At the end of surgery, desflurane and nitrous oxide were discontinued and the patients received 100% oxygen (7 L/min fresh gas flow).  Control: patients were breathing 100% oxygen with a fresh gas flow of 7 L/min. for 2 to 3 minutes. A standardized anaesthetic technique consisting of propofol for induction (2 mg/kg) followed by desflurane for maintenance of anaesthesia was used in all patients. The concentration of anaesthetic given to the patients was based on previously determined MAC values and adjusted to the patient's needs as clinically indicated with the objective to maintain the heart rate and blood pressure within 20% of the baseline values. They received a pre‐induction dose of fentanyl 2 μg/kg; additional doses of fentanyl 1 μg/kg were given if there were signs of inadequate anaesthesia (i.e. movement, swallowing, tearing, salivation) despite changes in inhalation concentration. Muscle relaxation for intubation was achieved by a single dose of vecuronium 0.1 mg/kg. Mechanical ventilation was instituted in all patients and ventilatory settings were adjusted to achieve normocapnia; the fresh gas flow was reduced to 2 L/ min during maintenance of anaesthesia. At the end of surgery, desflurane was discontinued and the patients received 100% oxygen (7 L/min fresh gas flow).
Outcomes	Secondary outcomes: Severe nausea and vomiting: vomiting requiring at least three doses of antiemetic medication within 24 hours of surgery.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Described as randomized but no further details.
Allocation concealment (selection bias)	Low risk	Sealed envelopes were used.
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	This was not reported.
Blinding of outcome assessment (detection bias)   Inhospital case fatality rate/length of stay	Unclear risk	—
Blinding of outcome assessment (detection bias)   Complications	Unclear risk	This was not reported.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No missing data.
Selective reporting (reporting bias)	Low risk	All outcomes described in methods section reported.
Other bias	High risk	Fewer than 50 participants per arm.

Study	Reason for exclusion
Antonini 1994	Not a RCT.
Atanassoff 1994	Not general anaesthesia.
Atassi 2005	Nitrous oxide used in the control group.
Barr 1999	Not a RCT.
Bronco 2010	Nitrous oxide used in the control group.
Castéra 2001	Not general anaesthesia.
Cheong 2000	Nitrous oxide used in the control group.
Divatia 1996	Not a RCT.
Dover 1994	Not a RCT.
Einarsson 1997	Nitrous oxide used in the control group.
Fredman 1998	Nitrous oxide used in the control group.
Gozdemir 2007	Nitrous oxide used in the control group.
Haessler 1993	Nitrous oxide used in the control group.
Haraguchi 1995	Not general anaesthesia.
Heath 1996	Not general anaesthesia.
Holst 1993	Nitrous oxide used in the control group.
Ishii 1994	Nitrous oxide used in the control group.
Jastak 1973	Non‐adults involvement.
Jellish 1996	Nitrous oxide used in the control group.
Johnson 1997	Non‐adults involvement.
Kryshtalskyj 1990	Not general anaesthesia.
Lim 1992	Non‐adults involvement.
Losasso 1992	Non‐adults involvement.
Masood 2002	Not general anaesthesia.
Morimoto 1997	Not a RCT.
Nightingale 1992	Non‐adults involvement.
Nishiyama 1998	Nitrous oxide used in the control group.
Ogg 1983	Non‐adults involvement.
Rocca 2000	Non‐adults involvement.
Saïssy 2000	Non‐adults involvement.
Simpson 1977	Nitrous oxide used in the control group.
Sinha 2006	Nitrous oxide used in the control group.
Smith 1993	Nitrous oxide used in the control group.
Taki 2003	Non‐adults involvement.
Towey 1979	Non‐adults involvement.
Van den Berg 1995	Non‐adults involvement.
Vari 2010	Nitrous oxide used in the control group.
Wesner 2005	Not a RCT.
Yamakage 2001	Nitrous oxide used in the control group.
Yang 2004	Nitrous oxide used in the control group.
Zuurmond 1986	Nitrous oxide used in the control group.

Methods	RCT
Participants	20 ASA 1 to 2 patients 18 to 60 years of age scheduled for orthopaedic surgery
Interventions	Intervention : intravenous combined with inhaled anaesthesia ventilated with 1.2 to 2.4 volume % isoflurane in nitrous oxide and oxygen Control : total intravenous anaesthesia ventilated with air and oxygen, FiO₂ 33%
Outcomes	Unknown
Notes

Methods	Controlled study
Participants	20 ASA 1 to 2 patients
Interventions	Intervention : thiopental and nitrous oxide in oxygen Control : propofol
Outcomes	Unknown
Notes

Methods	RCT
Participants	25 female patients during gynaecological laparotomies
Interventions	Intervention : sevoflurane in air and oxygen Control : sevoflurane in nitrous oxide and oxygen
Outcomes	Unknown
Notes

Methods	RCT
Participants	200 pregnant patients
Interventions	Intervention : thiopental‐nitrous oxide anaesthesia Control : total intravenous anaesthesia
Outcomes	Unknown
Notes	Only title available

PERMALINK

Nitrous oxide‐based techniques versus nitrous oxide‐free techniques for general anaesthesia

Rao Sun

Wen Qin Jia

Peng Zhang

KeHu Yang

Jin Hui Tian

Bin Ma

Yali Liu

Run H Jia

Xiao F Luo

Akira Kuriyama

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Intervention:

Control:

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

1.

Included studies

Excluded studies

Studies awaiting classification

Ongoing studies

Risk of bias in included studies

2.

3.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Summary of findings for the main comparison. Summary of findings’ table 1.

Summary of findings 2. 'Summary of findings' table 2.

Primary outcomes

1. Inhospital case fatality rate (number or proportion of deceased participants after a defined period following anaesthesia)

1.1. Analysis.

1.10. Analysis.

1.22. Analysis.

1.18. Analysis.

Secondary outcomes

1. Pulmonary complications