Nitrous Oxide for Colonoscopy

Abstract

Background

Colonoscopy is the gold standard investigation for large bowel disease. With the increase in demand, pressure is on clinics to shorten lengths of time per procedure in addition to maintaining high levels of patient safety. Analgesia has always been the mainstay of adequate pain relief, but it leads to prolonged recovery and lengths of hospital stay, in addition to increased risk of cardio‐respiratory side effects. N2O/O2 mixtures have been used for its effective analgesic effect and short half life and provides an alternative method of sedation for colonoscopy procedures.

Objectives

The primary objective was to compare the overall effectiveness of nitrous oxide mixtures to other types of pain relief used during colonoscopy procedures to provide adequate pain/discomfort relief.

The secondary objective was to compare between nitrous oxide and other types of pain relief with respect to hospitalisation/recovery time, side effects, patients and endoscopists satisfaction, and colonoscopy completion rates.

Search methods

The following electronic databases were searched: Cochrane Central Register of Controlled Trials (CENTRAL) on The Cochrane Library, MEDLINE (1966‐ present), EMBASE (1980 ‐ present), and the Internet (Google Scholar).

Selection criteria

Randomised controlled trials which compared nitrous oxide to placebo or active comparators for patients undergoing elective colonoscopic procedures. Patients with known underlying causes of pain/discomfort were excluded. 

Data collection and analysis

Seven randomised trials were included. Each trial compared a nitrous oxide/oxygen mixture to a placebo or sedation +‐ other analgesic drugs on patients undergoing elective colonoscopic procedures. The results of these studies were analysed and discussed.

Main results

There were a total of 547 patients included.

There were 257 patients randomised to receive the N2O/O2 mixture (7 studies), while 225 patients received some form of sedation with or without other analgesia (6 studies), and 65 patients received a placebo (3 studies).

Four studies showed that N2O/O2 is as good in controlling pain/discomfort as conventional methods, while one showed sedation was better and another study showed N2O/O2 was better.

Six of the studies showed that N2O/O2 groups had quicker recovery times and shorter lengths of hospital stays while one study showed that there was no difference between the two groups.

Two studies showed that N2O/O2 was safer while one reported that sedation was safer.

Authors' conclusions

Nitrous oxide is as efficient and safer than various pain relief methods used during colonoscopy procedures, but further trials are necessary.

Plain language summary

The use of Nitrous Oxide gas during Colonoscopy procedures

Colonoscopy is now the main investigation for large bowel illnesses, including cancer, but it can be an uncomfortable or painful procedure. To help reduce the pain/discomfort, endoscopists use narcotics or hypnotics with or without other analgesics, but this, in turn, comes with risks on the heart as well as on breathing. This review found that Nitrous Oxide gas works as good sedation, it leads to faster recover after the procedure and a short stay at the endoscopy unit, with little risk to breathing or the heart.

Background

Description of the condition

Colonoscopies have become an invaluable tool in diagnosing numerous large bowel and terminal ileal diseases and are more and more relied upon and considered to be the gold standard investigation for colorectal disorders (Maslekar 2009, Hafner 2007, Hamamoto 2002). A Colonoscopy, with its diagnostic and therapeutic indications, is one of the most frequently performed procedures in modern day medicine and the requests for it is on the rise (Maslekar 2009). This increase in requests for colonoscopies has even lead to a reduction in availability of highly skilled and experienced endoscopists, leaving other routine investigations such as flexible sigmoidoscopy at risk of not meeting the service needs, which led to the training of endoscopists who are non‐medically qualified to fill the gap left by colonoscopists (Gardiner 2009).

Colonoscopies are considered one of the most important investigations for large bowel pathology; they are relatively safe, but not without their complications (Hafner 2007, Maslekar 2009). Incidences of complications such as bleeding ranged between 1.6‐7% and bowel perforation ranging between 0.03‐0.19%, while pain and discomfort were the main complication ranging between 20‐41% depending on the severity (Hafner 2007). Colonoscopies are considered to be painful and a very distressing procedure for patients. One study reported that only 16.9% of 434 patients were willing to undergo the procedure without sedation (Early 1999). This pain had been contributed to numerous factors including stretching of the colonic wall and mesenteric attachments from looping of the instrument shaft, over insufflation, the degree of force applied to the shaft, patients' pain threshold, sex, age, body mass index, indications for the procedures, and even patients educational status (Shah 2002, Chung 2007, Early 1999, Dafnis 2005).

Controlling pain during a colonoscopy procedure enables the endoscopist to complete the procedure; a colonoscopy procedure is not considered fully complete unless the caecum is visualised, but most colonoscopists accept a 90% completion rate (Dafnis 2005, Chung 2007, Leung 2008). In order to reach a colonoscopy completion rate of 90%, the patient's pain and discomfort has to be controlled.

To try and alleviate or control the pain and discomfort during colonoscopies, physicians and surgeons have tried a wide range of procedures, instruments, techniques, medication, and even non conventional methods. Some of the methods used include different types of sedation, antispasmodic, sublingual hyoscyamine spray, patients controlled analgesia, nitrous oxide, variable stiffness colonoscopes, water instillation methods, electroacupuncture, music, and even positional manoeuvring of the patient (Hamamoto 2002, Tu 2006, Eckardt 1999, Maslekar 2009, Andrada 2004, Chaptini 2008, Fanti 2003, Hansel 2009, Saunders 1995, Yoshikawa 2008).

Though numerous methods have been tried, medication use remains the most widely used and is considered standard practice (Leung 2008, Ristikankare 1999, Early 1999). In the United States and Europe, patients receive an intravenous benzodiazepine, either alone or in combination with an opioid, for pain management (Leung 2008, Ristikankare 1999, Terruzzi 2001, Early 1999). Infact, trainee involvement in unsedated patients is considered impractical (Leung 2008). A recent observational study carried out in 10 European countries and Canada showed that 53% of patients received conscious sedation and 30% received deep sedation during their colonoscopy procedures (Hafner 2007). However, some centres prefer to give medications only on patient demand during the procedure, when pain and discomfort are no longer tolerated (Terruzzi 2001). In addition, some centres across the world do not use sedation on patients and have a variable colonoscopy completion rate ranging between 67‐95.3% (Chung 2007, Aljebreen 2004, Ristikankare 1999, Yoshikawa 2008, Leung 2008).

Medication given during procedures are not without side effects, sedatives, analgesics, narcotics and hypnotics all have cardiovascular and pulmonary complications. In fact, 20% of all patients undergoing an endoscopy with conscious sedation develop complications (Terruzzi 2001, Tu 2006, Early 1999, Eckardt 1999). Some of these side effects include respiratory depression, hypoxia, hypotension, paradoxical agitation, arrhythmias, occasionally respiratory arrest, all of which lead to an increase in morbidity and even mortality (Tu 2006, Terruzzi 2001, Eckardt 1999). These side effects lead to prolonged observation, monitoring, some patients requiring administration of other medications, and, therefore, a prolonged recovery time (Terruzzi 2001, Eckardt 1999).

To this end, numerous researchers have conducted studies comparing the Nirous Oxide to different medications in order find a safer alternative.

Description of the intervention

Nitrous Oxide (N₂O), was discovered by Joseph Priestley in 1793, but was first used as an anaesthetic in 1844 by Dr Horace Wells, a dentist who is now dubbed the "discoverer of anaesthesia" (Mennito 2006, Sullivan 2003)

N₂O, is a safe, tasteless, colourless gas which is rapidly absorbed and eliminated through the lungs (Mennito 2006, BOC3 2010, Sullivan 2003). It does not bind to haemoglobin or any other tissue, which is why it has a rapid onset and offset with minimal side effects (BOC2 2010, BOC3 2010, Sullivan 2003).

It produces both analgesic effects through interaction with the opioid receptors in the midbrain leading to activation of the descending inhibitory pathways which alters pain processing in the spinal cord and anxiolytic effect through interaction with the GABA‐A receptors (Mennito 2006, Sullivan 2003). An inhibitory neurotransmitter, gamma‐aminobutyric acid‐A (GABA‐A), inhibits pre‐synaptic cells from transmitting, therefore decreasing nervous system activity (Mennito 2006, Sullivan 2003).

N₂O is a strong analgesic, 20% N₂O is equal to 15mg of subcutaneous morphine (FRCA 2009). A mixture of 40‐70% N₂O added to oxygen (the remaining %), is used as an adjunct to inhalation and intravenous general anaesthesia, by which it helps start the process of anaesthetising the patient (Mennito 2006, FRCA 2009), while a lower percentage mixture of 25‐50% of N₂O can be used as an anxiolytic and analgesic during procedures (Mennito 2006, FRCA 2009, BOC 2010, BOC2 2010).

Benefits of using N₂O mixed with O₂ (BOC2 2010):

• Predictable and reliable analgesia,

• Effective pain relief and sedative effect without loss of consciousness,

• Rapid onset and offset, which can lead to savings through reduced treatment time and increased patient turnaround,

• Properties that help to reduce patient anxiety,

• Easy, self‐regulated administration,

• Minimal side‐effects

Though at therapeutic levels N₂O is considered safe, 100% N₂O can cause brain hypoxia which can lead to blackouts (Mennito 2006). However, there are some studies that suggest that N₂O can lead to a 2% drop in oxygen saturation, which, in a healthy individual, can be negligible, but is a problem in patients with an airway obstruction or pre‐existing cardiopulmonary disease such as pulmonary oedema and pulmonary hypertension, however, this can be avoided by breathing 100% oxygen for just a few minutes (Mennito 2006, Sullivan 2003). There are also studies that report that prolonged use of N₂O can lead to bone marrow suppression, by inactivating methionine synthase (a vitamin B12 dependent enzyme) it may case interference in the production of leukocytes and red blood cells, but this is only with prolonged use of N₂O for more than a total of 24 hours, or more frequently than every 4 days (Mennito 2006, Sullivan 2003, BOC2 2010). In addition, great caution should be taken when using N₂O on patients who are pregnant, especially early pregnancy, as there is a small risk of a miscarriage, though this is not proven (Mennito 2006). It must never be used if the patient has any condition where air is trapped in the body as it may lead to expansion (pressure effects) (BOC2 2010, Sullivan 2003):

1. pneumothorax

2. Intracranial air

3. abdominal distension or suspected intestinal obstruction

4. bullous emphysema

5. air embolism

6. middle ear procedures

7. following a recent dive/ decompression sickness

8. after recent eye surgery involving gas bubble insertion, due to the risk that the nitrous oxide will diffuse into the bubble, causing it to expand and eventually burst

How the intervention might work

Colonoscopy is a vital investigational tool used for investigating and sometimes treating large bowel disorders, and is used to keep an eye on recurrence of colorectal cancers, but the pain and discomfort that come with the procedure sometimes limit its completion (ASGE 2000, Tu 2006). Colonoscopists try to alleviate the patients pain and discomfort by giving them conscious sedation and strong analgesics, that can cause serious cardiovascular complications (Leung 2008, Ristikankare 1999, Terruzzi 2001, Tu 2006, Early 1999, Eckardt 1999).

As an inhaled gas that has an analgesic effect equivalent to morphine, but without the side effects, N₂O has been used instead of sedation methods and/or strong analgesics during medicinal and dental procedures since 1844 (Mennito 2006, Sullivan 2003). A recent study has reported that nitrous oxide with oxygen administered to the patient during a colonoscopy has better pain relief and a faster recovery time than sedation with analgesia (Maslekar 2009). However, Forbes et al, reported in their study that N₂O/O₂ is less effective than sedation with analgesia, but does allow for faster recovery (Forbes 2000). Using N₂O/O₂ instead of sedation with or without analgesia can greatly reduce the cardiopulmonary risks, in addition, it will reduce hopitalization and recovery time as well (Maslekar 2009, Forbes 2000, Leung 2008, Ristikankare 1999, Terruzzi 2001, Tu 2006, Early 1999, Eckardt 1999).

Why it is important to do this review

Use of N₂O instead of sedation and/or analgesia can lead to shorter hospital stays and recovery times with less risk of side effects, as well as a reduction in costs. Therefore, a thorough review of the literature and a meta‐analysis of the data is required to evaluate whether or not N₂O is as effective as sedation and/or analgesia to control pain during colonoscopies, results of which may lead to a change in colonoscopy practice worldwide.

Objectives

The primary objective was to compare the overall effectiveness of nitrous oxide mixtures to other types of pain relief used during colonoscopy procedures to provide adequate pain/discomfort relief.

The secondary objective was to compare the length of time of hospitalisation, the patients recovery time, procedure length of time, the procedure completeness, any colonoscopic complications, any side effects from the medication used, any difference with the procedure itself with regards to operability when nitrous oxide was used compared to other methods, and the patients' as well as the colonoscopists' assessment of the procedure.

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials comparing nitrous oxide to placebo or active comparators was considered for inclusion. 

Types of participants

Participants were patients undergoing elective colonoscopic procedures.

Patients undergoing emergency colonoscopy, or who have a history of pneumothorax or severe asthma, previous colonic resection, intolerance or allergy to any of the drugs, patients taking opiates, or patients that are undergoing upper gastrointestinal endoscopy and colonoscopy simultaneously were excluded.

Types of interventions

Patients receiving Nitrous oxide were compared to those receiving a placebo, an analgesia, sedation, acupuncture, narcotics, hypnotics or any other method of pain relief.

Types of outcome measures

Primary outcomes

The primary outcome will be adequate pain relief during the procedure. Adequate pain relief can be measured by comparing pain scores of patients who had N2O or sedation by pain questionnaires such as the Visual Analogue Scale (VAS), by numerical or picture representation of the pain experienced during and after the procedure, or by any alternative to means to measure the pain.

Secondary outcomes

Secondary outcomes included length of hospitalisation, patient recovery time, procedure length of time, the procedure completeness, any colonoscopic complications, any side effects from the medication used, and any operative difficulties.

Search methods for identification of studies

Electronic searches

The following electronic databases were searched: Cochrane Central Register of Controlled Trials (CENTRAL) on The Cochrane Library, MEDLINE (1966‐ present), EMBASE (1980 ‐ present), and the Internet (Google Scholar).

The following search strategy were used to identify relevant randomised trials:

#1 colonoscopy OR endoscopy

#2 Nitrous Oxide OR NO OR N2O OR Entonox

#3 Randomized Control Trials

#3 #1 AND #2

#4 #1 AND #2 AND #3

Searching other resources

#1 Referencing searching: The references of all identified studies were inspected for more trials.

#2 Personal contact: The contact author of each included study were contacted for data of their study or information regarding unpublished trials.

#3 Drug company: Producers of nitrous oxide including medication (such as Entonox (a registered trademark of BOC)) were contacted for additional data and/or information.

Data collection and analysis

Selection of studies

Three authors (OMA, TA, and SASC) independently assessed relevant titles and abstracts identified by literature search against the predefined inclusion criteria. Full papers of the titles and abstracts selected were further assessed for inclusion or exclusion. Any disagreement among the authors was resolved by consensus or by consultation with a third party (RLN).

Data extraction and management

Four authors (OMA, TA, SASC, and PJM) independently extracted data from the included studies using a predetermined data extraction form. Any disagreement was resolved by consensus or by consultation with a third party (RLN).  If needed the authors of the primary studies were contacted for clarification of data and/or additional information.

Assessment of risk of bias in included studies

The five authors (OMA, TA, SASC, PJM, and RLN) were assessed the risk of bias as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008).

Factors assessed include:

1) Sequence generation (i.e. was the allocation sequence adequately generated?);

2) Allocation sequence concealment (i.e. was allocation adequately concealed?);

3) Blinding (i.e. was knowledge of the allocated intervention adequately prevented during the study?);

4) Incomplete outcome data (i.e. were incomplete outcome data adequately addressed?);

5) Selective outcome reporting (i.e. are reports of the study free of suggestion of selective outcome reporting?); and 

6) Other potential sources of bias (i.e. was the study apparently free of other problems that could put it at a high risk of bias?).

A judgement of 'Yes' indicates low risk of bias, 'No' indicates high risk of bias, and 'Unclear' indicates unclear or unknown risk of bias.

Measures of treatment effect

Review Manager Analyses (RevMan 5.0.23 Java 6) was used to analyse data. Analyses was performed using the intention to treat principle. Data from individual trials was combined for meta‐analysis when the interventions and controls are similar.  Dichotomous data was calculated as relative risk (RR) with 95% confidence intervals (95% CI). The weighed mean difference (WMD) with 95% CI was calculated for continuous outcomes. 

Unit of analysis issues

In addition to simple parallel group designs, variations such as cluster‐randomisation and cross‐over trials was analysed.

Dealing with missing data

The authors of papers with Missing data (studies, outcomes, summary data, individuals, or study level characteristics) were contacted for request of the missing data.

Where the data can not be requested or the authors of the corresponding papers unavailable the study was excluded.

Assessment of heterogeneity

A fixed effects model was used unless statistically significant heterogeneity exists between the studies, there in a random effects model was employed. A chi‐square test was used to assess heterogeneity (a P value of 0.10 will be regarded as statistically significant). If statistically significant heterogeneity is identified, the included studies was examined for sources of both clinical and methodological heterogeneity. 

Assessment of reporting biases

The possibility of a publication bias was investigated through the construction of funnel plots (trial effects versus trial size).

Data synthesis

A fixed effects model was used unless statistically significant heterogeneity exists between the studies. A random effects model was employed if heterogeneity exists.

Subgroup analysis and investigation of heterogeneity

We intended to explore the following potential sources of heterogeneity using sub‐group analyses or meta‐regression:

1. Intervention ‐ Nitrous oxide and different control;

2. Different types of nitrous oxide mixture;

3. Different duration of treatment;

4. The grade or level of expertise of the colonoscopists;

5. Indications for the procedure.

Sensitivity analysis

Sensitivity analyses was not conducted as a meta‐analysis could not be done.

Results

Description of studies

See Characteristics of included studies

Results of the search

The article selection process is displayed in Figure. The literature search yielded a total of 304 articles. Screening of the Titles and/or Abstracts of the articles found that 278 articles did not meet the inclusion criteria and therefore were excluded. Full manuscripts of 26 articles were reviewed of which 7 were included (see table of Characteristics of included studies) and 19 were excluded (see table of Characteristics of excluded studies).

1.

Flow chart of the article selection process

Included studies

Seven randomised control trials met the inclusion criteria (see table of characteristics of Included Studies). All of the randomised control trials compared Nitrous Oxide (N2O) to narcotics or a hypnotic in patients undergoing elective colonoscopy procedures (Forbes 2000, Lindblom 1994, Maslekar 2009 BJS, Maslekar 2009 CD, Notini‐Gudmarsson 1996,Saunders 1994, Trojan 1997) and only one study compared N2O to a narcotic and to a placebo (Saunders 1994).

A statistical meta‐analysis was not possible on the included studies because the results of the data could not be pooled or compared for all the studies, therefore a detailed description of each paper will be discussed, a summary of the results are in Table. The main reasons why is that only two studies compared the same group of analgesics to N2O (Saunders 1994 and Trojan 1997) while the other RCTs compared different combinations of pain relief. In addition, it was not possible to statistically analyse the results as some of the data given in the RCTs were in medians, others in means with some data not providing ranges or standard error. As shown by Michiels 2005, it would not be sufficient to use medians for meta‐analyses as you will not only have an over or under‐estimation of the overall hazard ration, but you will also lose statistical power (Michiels 2005) therefore we will need individual patient data, so you can determine the spread of the data, but we were unable to get the data. In short medians are not good for meta‐analyses because you do not know how the data is distributed above and below the median even if the range is known. Hence, we thought it prudent to discuss the results of each RCT individually, however we have conducted a meta‐analyses of the Caecal intubation rates which was the only data that can be pooled for comparison (Figure).

1. Study Measurments of Results.

Study ................................	Total N= ...............	Used N2O ................	Comparators .....................	Patients Overall Assessment .............................................	Colonoscopist Assessment ...........................................	Pain/Discomfort ...............................	Caecal Intubation Rates ......................................	Time to Caecum ..........................	Total Colonoscopy Time ........................................	Time to Recovery ............................
Maslekar 2009BJS	131	Yes	Entonox vs. Midazolam & Fentanyl	+ Median (range)	+ Median (range)	+ Mean (range)	+	+ Median (range)	+  MedIan (range)	+ Mean (range)
Maslekar 2009CD	100	Yes	Entonox vs. Propofol	+ Median (range)	+ Median (range)	+ Median (IQR)	+	+ Mean (range)	+  mean (range)	+ Mean (range)
Forbes 2000	102	Yes	Entonox vs. Midazolam & Meperidine	+ Median		+ Mean	+	+ Mean	+  Mean	+ Median (range)
Trojan 1997	37	Yes	Entonox vs. Midazolam & Pethidine vs. no sedation	‐	‐	‐	+	‐	+  Mean	‐
Notini‐Gudmarsson 1996	38	Yes	Medimix vs. Pethidine	+ Median +‐ SD	+ Median	+ Mean +‐ SD	+	‐	+ Median (range & +‐ SD)	+ Mean (range)
Lindblom 1994	50	Yes	Entonox vs. Placebo	‐	‐	+ Median (range)	+	‐	+  Median (range)	+ Median (range)
Saunders 1994	89	Yes	Entonox vs. Midazolam & Pethidine vs. Placebo	‐	‐	+ Median	+	+  Median	‐	+ Median

2.

Forest plot of comparison: 1 Caecal Intubation Rate, outcome: 1.1 Intubation Rate.

All of the studies were carried out on adults with an approximately equal male: female ratio of 1.1:1. 
 Five studies used different types of medication; Maslekar 2009 BJS used IV midazolam (2mg/ml) and fentanyl (0.075‐0.1mg), Maslekar 2009 CD used IV propofol (through a patient controlled pump), Forbes 2000 used IV midazolam (0.06mg/kg) and meperidine (0.76mg/kg), Lindblom 1994 used IV ketobemidone hydrochloride (2.5mg) and midazolam (2.5mg), and Notini‐Gudmarsson 1996 used IM Pethidine hydrochloride (1mg/kg). While two studies (Trojan 1997 and Saunders 1994) both used IV midazolam (2.5mg) and pethidine (25‐50mg). 
 Six studies used Entonox (Forbes 2000, Lindblom 1994, Maslekar 2009 BJS, Maslekar 2009 CD, Saunders 1994, Trojan 1997), while only Notini‐Gudmarsson 1996 used Medimix. Both Entonox and Medimix contain Nitrous Oxide mixed with Oxygen in a 50:50 ratio. 
 Two of the studies used a premedication hyoscine butylbromide (20mg) on all of their patients (Saunders 1994 and Trojan 1997).

See Table

2. Summery of Results.

Study .......................	Results Summery ...........................................................................................................................................................................................................................................	Study Conclusion of outcomes ....................................................................................
Maslekar 2009BJS	Pain score: mean VAS score 16·7 versus 40. 1; P = 0·001 depth of sedation: median (range) 4 (4‐5) and 3 (3‐4); P = 0.163 Caecal intubation rates: 94% vs. 92%; P = 0.513 time to reach caecum: median time to reach caecum minutes ( range): 18 (6‐24) vs. 20 (5‐23); P = 0.812 total completion time: median minutes (range): 26.5 (18‐47.5) vs. 31.9 (20.4‐51.8); P = 0.761 psychomotor recovery: immediate %(range):  92.2 (89.5‐96) vs. 0 15min.s after procedure: 94·3 (92·5?100) vs. 68·4 (60·8?71·1); P = 0·001; at Discharge: 100 vs. 87·3 (82·0?94·6); P = 0·003 time to discharge: mean minutes (range): 28 (12?38) vs. 51 (34?76); P = 0·001 side effects: one patient who developed nausea in the sedation group patients’ overall satisfaction score: median score on a VAS (range): 96 vs. 89; P = 0·024 endoscopists’ overall satisfaction score: median score on a VAS (range): 95 (90‐100) vs. 90 (70‐100); P = 0.061 nurses’ overall satisfaction score: median score on a VAS (range): 95 (30‐100) vs. 90 (20‐100); P = 0.168	Entonox is more effective than Midazolam & Fentanyl in providing pain relief and led to a quicker recovery.
Maslekar 2009CD	Pain score: median scores at 15 minutes 15.78 (interquarterly range (IQR): 12‐20) vs. 16.54 (IQR: 10‐20), at discharge 15.38 (IQR: 14‐20) vs. 17.31 (IQR: 10‐20), 24 hours post procedure 16.14 (IQR: 12‐21) vs. 17.89 (IQR:10‐20) depth of sedation: P = 0.091 manoeuvrability during the procedure: P = 0.2 Caecal intubation rates: 96% vs. 98%; P = 0.55 time to reach caecum: mean minutes (range) 13 (10‐16.25) vs. 14 (12‐23.25); P = 0.18 total completion time: mean minutes (range)22.6 (18‐28) vs. 20.8 (17‐23); P = 0.09 psychomotor recovery: immediate % (range):  92 (89.5‐96) vs. 90 (84‐92); P = 0.79; 15min.s after procedure: 99 (98.5?100) vs. 96 (91?99); P = 0.08; at Discharge: 100 vs. 97 (94‐100);P = 0.07 time to discharge: mean minutes (range): 27.86 (22?30.5) vs. 28.08 (23‐32); P = 0.86 patients’ overall satisfaction score: median (range) VAS: vs. 96 (94‐98); P=0.1 nurses’ overall satisfaction score: median (range) VAS: 95.8 (93.4‐98) vs. 97.24 (95‐99); P=0.34 endoscopists’ overall satisfaction score: median (range) VAS: 96 (95‐98) vs. 98 (96‐100); P=0.26	Entonox & Propofol provide the same effectiveness in providing pain relief.
Forbes 2000	Caecal intubation: 100% vs. 98% time to caecum: mean 9.8 vs. 7.5; P=0.06 time to completion: mean 17.4 vs. 16.6; P=0.6 difficulty of the procedure: 26 vs. 17; P=0.18 tolerance of the procedure: 21 vs. 5; P<0.0001 pain: 33 vs. 3; P<0.0001 patients’ satisfaction: 13 vs. <0.01 recovery time:  median (range): 30 (30‐45) vs. 60 (30‐110); P=0.0001 side effects: 27% vs. 11%	Entonox is less effective than Midazolam & Meperidine in controlling pain but led to a quicker recovery.
Trojan 1997	Duration of procedure: mean minutes (+‐ SD): 28+‐3 vs. 30+‐3 vs. 22+‐3 Difficulty: 42% vs. 13% vs. 0% psychomotor testing: tracking test: immediately after P=0.03, at 15 min P=0.04, at 30 minutes P=0.02, and at 45 min P=0.01). manual dexterity test: immediately (P=0.01), at 15 min (P=0.04) and at 30 min (P=0.04) letter cancellation test (P=0.04) multiple choice reaction time: no significant difference	Entonox & no sedation led to a quicker recovery than Midazolam & Pethidine
Notini‐Gudmarsson 1996	Pain: end of the procedure: 44+‐26 vs. 58+‐27; just before discharge: 41+‐30 vs. 54+‐29 Nausea: after the procedure 4+‐8 vs. 15+‐21; P<0.05 and just before discharge 3+‐6 vs. 21+‐29 P<0.05 completion time: 17+‐6 vs. 16+‐7 time to discharge: 56+‐31 minutes vs. 96+‐60 minutes; P<0.05 patients’ experience: end of the procedure 19+‐19 vs. 31+‐25; P<0.05 and just before discharge 16+‐15 vs. 30+‐27;P=0.064 colonoscopists’ evaluation: 6.4 vs. 6.2 side effects: pulse rate change: 8+‐1 vs. 18+‐2; P<0.01 and O2 % saturation during 98+‐0.8 vs. 95+‐2.9; P<0.01 and 5 minutes after: 98+‐1.1 vs. 95+‐2.2; P<0.01	Entonox & Pethidine provide the same effectiveness in providing pain relief but a quicker recovery.
Lindblom 1994	Discomfort: median VAS score (range): 51 (28‐65) vs. 54 (24‐74); P=0.64 recollection tests: median score (range): 8.5 (8‐9) vs. 7 (6‐8.5); P=0.02 recovery time: median minutes (range): 0 (0‐5) vs. 37.5 (10‐75); P=0.0001 procedure time: median minutes (range): 47 (33‐53) vs. 45 (35‐58); P=0.67 extra medication needed: patient no 1 vs. 5; P=0.09	Entonox is more effective than the placebo in providing pain relief and led to a quicker recovery.
Saunders 1994	Pain: median VAS: 10 vs. 15; P=0.85 patients’ experience with depth of sedation: 20 vs. 56; P=0.009 or sedation vs. placebo group median 5; P=<0.0001 duration of stay: median minutes 32 vs. 60; P=<0.001 or sedation vs. placebo group median 36; P=0.002 side effects: O2 saturation <90% for more than 1minute: 0 vs. 6 patients; P=0.01, hypotensive episodes (drop in blood pressure of >20 mm Hg to <100 mm Hg associated with sweating and pallor): 2 vs. 4 patients and 4 in the placebo group	Entonox is more effective than Midazolam & Pethidine in providing pain relief and led to a quicker recovery.

Maslekar 2009 BJS conducted a randomised controlled trial on 131 patients comparing Entonox and midazolam‐fentanyl with 65 patients in the Entonox group and 66 in the sedation group. The authors stated that the two groups had similar demographics and ASA grades. No difference was found with the pre‐procedure anxiety score between the two groups 7·5 (range 3–10) versus 6.0 (range 2–12); P = 0.143). Patients in the Entonox groups recorded significantly less Pain than those in the sedation group (mean VAS score 16·7 versus 40. 1; P = 0·001; 95 percent confidence interval (c.i.) of the difference 21·5 to 29·1). The median depth of sedation using the MOAAS was 4 (4‐5) and 3 (3‐4) (P = 0.163). The Caecal intubation rates were 94% in the Entonox group vs. 92% for the sedation group (P = 0.513) with no difference between the two groups with regards to time to reach caecum or total completion time (median time to reach caecum minutes; range): 18; 6‐24 vs. 20; 5‐23 (P = 0.812) and total completion time median (minutes; range): 26.5; 18‐47.5 vs. 31.9; 20.4‐51.8 (P = 0.761). There was significant difference between the two groups with regards to psychomotor recovery and time to discharge, with the Entonox group having quicker functional recovery immediately and 15 minutes after the procedure in addition to significant difference to mean time to discharge (immediate %; range):  92.2; 89.5‐96 vs. 0; 15 minutes after procedure: 94·3; 92·5–100 vs. 68·4; 60·8–71·1 (P = 0·001); at Discharge: 100 vs 87·3; 82·0–94·6 (P = 0·003); mean time to discharge (minutes; range): 28; 12–38 vs. 51; 34–76 (P = 0·001)). The Patients overall Satisfaction Score was significantly higher in the Entonox group in both before receiving the colonoscopy results, i.e. after the procedure and 24 hours after the procedure (96 versus 89; P = 0·024; 95% c.i. of the difference 4 to 12). The Endoscopists’ and Nurses Satisfaction Score of the procedure showed no difference between the two groups (median score on a VAS: 95; 90‐100 vs. 90; 70‐100 (P = 0.061) and for nurses scores: 95; 30‐100 vs. 90; 20‐100 (P = 0.168)).

Maslekar 2009 CD conducted a randomised controlled trial on 100 patients comparing Entonox and propofol with 50 patients in the Entonox group and 50 in the sedation group. The authors reported that there were no differences between the Entonox group and the Propofol group with regards to baseline characteristics. The two had similar pre‐procedure anxiety scores (7.5 vs. 8.4; P = 0.1). Their were no statistical significances between the two groups with regards to the Pain score, assessed via VAS. The median pain scores at 15 minutes after the procedure in the Entonox group was 15.78 (interquarterly range (IQR: 12‐20) vs. 16.54 (IQR: 10‐20) in the Propofol group (95% CI of difference in means ‐0.93, 5.06; pre‐determined equivalence margin ‐15, 15), at discharge the scores were 15.38 (IQR: 14‐20) vs. 17.31 (IQR: 10‐20) (95% CI of difference in means ‐0.89, 5.02; pre‐determined equivalence margin ‐15, 15), and 24 hours post procedure the scores were 16.14 (IAR: 12‐21) vs. 17.89 (IQR:10‐20) (95% CI of difference in means ‐0.88, 5.03; pre‐determined equivalence margin ‐15, 15). There was no statistical significance between the Entonox and the Propofol groups in regards to Depth of Sedation (P = 0.091) and manoeuvrability during the procedure (P = 0.2). In addition there were no differences between the two groups in Caecal Intubation percentage (96% vs. 98%; P = 0.55; 95% CI of the difference in means ‐0.56,; predetermined equivalence interval ‐5, 5), mean time to reach Caecum (minutes)(13 (10‐16.25) vs. 14 (12‐23.25); P = 0.18; 95% CI of the difference in means ‐1.82‐3.82,; predetermined equivalence interval ‐5, 5), and mean total time to completion (minutes)(22.6 (18‐28) vs. 20.8 (17‐23); P = 0.09; 95% CI of the difference in means ‐0.83‐4.36; predetermined equivalence interval ‐5, 5). There was no significant difference between the two groups with regards to psychomotor recovery and time to discharge, with the functional recovery immediately, 15 minutes after the procedure or at discharge (immediate (%; range):  92; 89.5‐96 vs. 90; 84‐92 (P = 0.79); 15min.s after procedure: 99; 98.5–100 vs. 96; 91–99 (P = 0.08); at Discharge: 100 vs. 97; 94‐100 (P = 0.07); mean time to discharge (minutes; range): 27.86; 22–30.5 vs. 28.08; 23‐32 (P = 0.86)). The Endoscopists, Nurses, and Patients satisfaction overall with the procedure showed no difference between the Entonox group and the Propofol group (median (range) VAS: Endoscopist: 96 (95‐98) vs. 98 (96‐100); P=0.26, Nurse: 95.8 (93.4‐98) vs. 97.24 (95‐99); P=0.34, Patients: 94 vs. 96 (94‐98); P=0.1).The Entonox group required a median of 2‐4 hours to resume normal activities compared to 6 hours in the Propofol group (P=0.02), where as 54% of the Entonox groups required additional sleep compared to 96% in the Propofol group (P=0.03).

Forbes 2000 conducted a RCT on 102 patients to receive either Entonox (n=56) or sedation (n=46) with midazolam and meperidine. The author mentions that there were more women randomised to receive Entonox rather than sedation but this was statistically not significant (P=0.23), however, men had a lower anxiety score than women (P=0.008). Caecal intubation was 100% in the sedation group compared to 98% in the Entonox group with no significant difference between the two groups with regards to mean time to caecum (Entonox: 9.8 vs. Sedation 7.5(P=0.06) or mean time to completion (Entonox: 17.4 vs. Sedation 16.6 (P=0.6). Although there was no significant difference between the endoscopists assessment of the difficulty of the procedure between the two groups (VAS: Entonox: 26 vs. sedation: 17 (P=0.18), the endoscopists perception was that the sedation group tolerated the procedure better (median tolerance on VAS: Entonox: 25 vs. Sedation 12(P=0.0003). The patients assessment of the procedure revealed that patients in the Entonox group recalled more pain compared to those in the sedation group (median VAS: Entonox 33 vs. sedation:3 (P<0.0001)), they had less tolerance of the procedure (Entonox 21 vs. sedation:5 (P<0.0001)), were less satisfied with the procedure (Entonox 13 vs. sedation:6 (P<0.01)), and would less readily have the procedure repeated (Entonox 6 vs. sedation:2 (P<0.04)). In addition more patients in the Entonox group found the procedure more unpleasant than expected (Entonox 27% vs. sedation: 0% (P<0.0001)). Patients in the Entonox group had a faster time to recovery than those in the sedation group (median (range): Entonox: 30 minutes (30‐45) vs. sedation: 60 (30‐110) (P=0.0001)).

Trojan 1997 conducted a RCT on 37 patients to receive either nitrous oxide/oxygen mixture (n=12), pethidine and midazolam (n=15), or no sedation (n=10). The demographics were similar in all three groups. The procedures was completed in all the patients. The mean duration of the procedure was 28+/‐3minutes for the N2O/O2 group, 30+/‐3 minutes for the sedation group and 22+/‐3 minutes for the unsedated group. The procedure was deemed difficult in 42% of the patients in the N2O/O2 group compared to 13% in the sedation group and 0% in the unsedated group, no statistics were provided. The first psychomotor test was done sooner than in the N2O/O2 and unsedated groups, however when compared to the sedation group this did not reach statistical significance (median minutes (range): 8 (3‐25) vs. 16 (3‐50) (P=0.11 and 0.07)).  There were four psychomotor tests conducted after the procedure. The tracking test showed that the N2O/O2 group had significant improvement when compared to baseline results (immediately after the procedure P=0.03, at 15 min P=0.04, at 30 minutes P=0.02, and at 45 min P=0.01). In the sedation group on 45 min after the procedure the patients showed improvements (P=0.02). There was no difference between the N2O/O2 and the unsedated groups. When comparing the sedation group to the unsedated group there was significant difference both immediately (P=0.04) and at 15 min (P=0.02). There were no significant differences in the multiple choice reaction time between the three groups. With regards to the manual dexterity test, the N2O/O2 and unsedated groups showed no prolonged time to complete the test, whereas the sedation group showed significant prolongation immediately (P=0.01), at 15 min (P=0.04) and at 30 min (P=0.04). The letter cancellation test revealed that there was impairment in the sedation group (P=0.04) and there was a significant difference when compared to the unsedated group (P=0.04) otherwise the tests were not impaired and no differences were found.

Notini‐Gudmarsson 1996 conducted a RCT on 38 patients to receive either N2O/O2 (n=10) or pethidine (n=19).  There were no differences between the groups with regards to their demographics. Total colonoscopy was achieved in 85% in both groups. No significant difference was found between the two groups with regards to pain score using a VAS at the end of the procedure and just before discharge (Medimix: 44+‐26 vs. pethidine: 58+‐27 and Medimix: 41+‐30 vs. Pethidine: 54+‐29). However, the Medimix group had significantly less nausea scores than the pethidine group both just after the procedure and just before discharge (4+‐8 vs. 15+‐21 (P<0.05) and 3+‐6 vs. 21+‐29 (P<0.05)).  There were no statistical differences with respect to the procedure completion time (Medimix: 17+‐6 vs. pethidine: 16+‐7) or the colonoscopists evaluation score (Medimix: 6.4 vs. pethidine: 6.2), but no statistical information was provided. Time to discharge was also significantly quicker in the Medimix group (56+‐31 minutes vs. 96+‐60min (P<0.05)). The total patient experience of the procedure showed that patients in the Medimix group had better overall experience than those in the sedation group at the end of the procedure (19+‐19 vs. 31+‐25 (P<0.05)), however, there were no differences between the experiences of the two groups just before discharge (16+‐15 vs. 30+‐27 (P=0.064)).

Lindblom 1994 conducted a RCT on 50 patients to receive either Entonox (n=25) or air via face mask (KHM) as a placebo (n=25). There were no differences in the demographics between the two groups with regards to sex (Male/Female: 14/11 in the Entonox group; 12/13 in the KHM group (P=0.38)), age (median (range): 46 (38‐56) vs. 40 (30‐39) (P= 0.34), or previous colonoscopy experience (17 vs. 13 (P= 0.19)). There were no significant differences between the two groups with regards to overall discomfort (median VAS score (range): Entonox group: 51 (28‐65) vs. KHM group: 54 (24‐74) (P=0.64)). There were also no differences between the groups with regards to pre‐procedure recollection test (median score (range): Entonox group: 9 (8‐9) vs. KHM group: 9 (8‐10)  (P=0.97)), however, a significant difference was found in the results of the post‐procedure recollection test (median score (range): Entonox group: 8.5 (8‐9) vs. KHM group: 7 (6‐8.5) (P=0.02) favouring the Entonox, i.e. Entonox groups had better recollection test scores. In addition, the Entonox group had shorter time to recovery compared to the KHM group (median minutes (range): Entonox group: 0 (0‐5) vs. KHM group: 37.5 (10‐75) (P=0.0001), but no difference was found in the total colonoscopy time (median minutes (range): Entonox group: 47 (33‐53) vs. KHM group: 45 (35‐58) (P=0.67)). There were also fewer patients in Entonox group who required extra medication than the KHM group (1 vs. 5 (P=0.09)

Saunders 1994 conducted a RCT on 89 patients to receive either IV sedation (n=29), N2O/O2 (n=30), or a placebo (n=30). The author states the baseline demographics for each of the groups were similar. There were no differences between the N2O group and the sedation group with regards to Pain (median VAS: N2O: 10 vs. Sedation: 15 (P=0.85)). However, both groups scored less pain than the placebo group (median VAS: 30) (P=0.05 for each comparison). Patients experience more sedation in the sedation group (median 20) compared to the N2O group (median 20) (P=0.009) or even the placebo group (median 5) (P=<0.0001), but no difference in experience of sedation was found between the N2O and the placebo groups (P=0.06). There was significantly longer duration of stay within the recovery area in the sedation group (median 60 minutes) compared to the N2O group (median 32 minutes) (P=<0.001) and the placebo group (median 36 minutes) (P=0.002). More patients developed oxygen desaturation in the sedation group (n=6) than in the N2O group (n=0) (P=0.01) and no significant difference was found between the sedation and N2O groups with regards to headaches experienced (P=0.2).

Maslekar 2009 BJS only reported one patient who developed nausea in the sedation group, while Notini‐Gudmarsson 1996 reported that the sedation group had a significantly higher rise in pulse rate compared to the Medimix group (Medimix: 8+‐1 vs. sedation: 18+‐2 (P<0.01)) in addition to significant difference in reduction O2 saturation during and 5 minutes after the procedure (During: Medimix: 98%+‐0.8 vs. sedation: 95%+‐2.9 (P<0.01); 5 minutes after: 98%+‐1.1 vs. 95%+‐2.2 (P<0.01)) with no differences in baseline saturation between the two groups.

While Saunders 1994 reported that 6/29 patients in the sedation group had O2 saturation <90% for more than 1minute compared to 0 in the Entonox group (P=0.01), in addition to 4/29 patients developing significant hypotensive episodes (drop in blood pressure of >20 mm Hg to <100 mm Hg associated with sweating and pallor) compared to 2/30 in the Entonox group and 4/30 in the placebo group, but these did not reach statistical significance. However, Forbes 2000 reported that adverse events occurred more commonly in the Entonox group (27% vs. 11%) but these were not clinically serious side effects.

Excluded studies

Eight of the nineteen studies were duplicates and removed the studies that were assessed. Eleven manuscripts were excluded from the final analysis (see table of characteristics of Excluded Studies). 
 Eight studies were found not to be randomised control trials (Eckardt 1999, Fink 1998, Maslekar 2006, Maslekar 2007, Michaud 1999, Phillips 1995, Trojan 1998, Trojan 1998GE). While 3 randomised trials (Ristikankare 1999, Terruzzi 2001, Theodorou 2001) compared different sedation methods but did not use N2O/O2 as one of the comparisons, therefore were excluded.

Risk of bias in included studies

See Description of studies, Figure, Figure, and Figure.

3.

Methodological quality graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.

4.

Methodological quality summary: review authors' judgements about each methodological quality item for each included study.

5.

Funnel plot of comparison: 1 Caecal Intubation Rate, outcome: 1.1 Intubation Rate.

Allocation

The sequence generation was unclear in 3 of the studies (Trojan 1997, Notini‐Gudmarsson 1996, and Forbes 2000). While, 4 studies had adequate sequence generation (Maslekar 2009 BJS, Maslekar 2009 CD, Lindblom 1994, Saunders 1994). 
 The allocation concealment was unclear in 4 of the studies (Trojan 1997, Notini‐Gudmarsson 1996, Forbes 2000, and Saunders 1994). While, 3 studies had adequate concealment (Maslekar 2009 BJS, Maslekar 2009 CD, and Lindblom 1994)

Blinding

Two studies have described themselves as double‐blinded trials (Saunders 1994 and Lindblom 1994). 
 Two studies mentioned that neither the patients or the investigators were blinded, therefore no blinding was done (Maslekar 2009 BJS and Maslekar 2009 CD). 
 The blinding was unclear in the remaining 3 studies (Forbes 2000, Trojan 1997, Notini‐Gudmarsson 1996).

Incomplete outcome data

All seven trials had no patients either drop out or withdraw from the studies. All the patients enrolled in the trials had their data included into the final results of each article and there was no missing or incomplete outcome data found.

Selective reporting

There were no indications of selective reporting in any of the seven trials.

Other potential sources of bias

One study had no other potential source of bias identified (Lindblom 1994). 
 Four studies were deemed to have a potential source of bias because the trials were not blinded or blinding was not clear (Maslekar 2009 BJS, Maslekar 2009 CD, Trojan 1997, Notini‐Gudmarsson 1996). 
 Three studies were unclear on how the allocation sequence was generated (Forbes 2000, Trojan 1997, and Notini‐Gudmarsson 1996) and 4 studies were unclear on how the allocation concealment was done (Saunders 1994, Maslekar 2009 BJS, Maslekar 2009 CD, Trojan 1997, Notini‐Gudmarsson 1996), therefore were considered to have a potential source of bias. 
 Two studies reported the endoscopists opinion of the patients tolerability of the procedure, which is subjective and can be considered a potential source of bias (Trojan 1997 and Forbes 2000).

Effects of interventions

(Note: Statistics will not be repeated as they are presented in the results section)

With regards to the primary objective of this review, out of the seven included articles only one article showed that N2O/O2 was better than the sedation in controlling the pain/discomfort (Maslekar 2009 BJS) and only one article reported that the sedation group experienced more pain (Forbes 2000), while four of the articles reported that there were no differences between the N2O/O2 and sedation groups and that N2O/O2 was as good in controlling the pain/discomfort as sedation +‐ analgesia/opiate (Maslekar 2009 CD, Notini‐Gudmarsson 1996, Lindblom 1994, Saunders 1994).

Concerning the secondary objectives:

There was no difference between the N2O/O2 and sedation groups with regards to Caecal intubation rates (Maslekar 2009 BJS, Maslekar 2009 CD, Forbes 2000,Trojan 1997, Notini‐Gudmarsson 1996).

No difference was found with regards to time needed to reach the caecum (Maslekar 2009 BJS, Maslekar 2009 CD, Forbes 2000) or total time needed to complete the colonoscopy procedure (Maslekar 2009 BJS, Maslekar 2009 CD, Forbes 2000,Trojan 1997, Notini‐Gudmarsson 1996, Lindblom 1994).

With regards to length of time required for recovery and/or discharging after the procedure, six of the seven included studies showed that the N2O/O2 groups were significantly quicker to recover or required less time to recover (Maslekar 2009 BJS, Forbes 2000, Trojan 1997, Notini‐Gudmarsson 1996, Lindblom 1994, Saunders 1994), while Maslekar 2009 CD reported that there were no differences between the two groups, N2O/O2 and sedation.

With regards to the overall patients experience of the procedure, one study reported that there was less satisfaction with the procedure in the N2O/O2 group than the sedation group (Forbes 2000). However, two other studies showed that there was significantly higher satisfaction with the procedure in the N2O/O2 groups rather than the sedation groups (Maslekar 2009 BJS and Notini‐Gudmarsson 1996). While Maslekar 2009 CD reported that there was no difference between the two groups.

With regards to the overall endoscopists experience of the procedure, four studies showed no difference between N2O/O2 and sedation (Maslekar 2009 BJS, Maslekar 2009 CD, Forbes 2000, Notini‐Gudmarsson 1996), while Trojan 1997 reported that there was more difficulty in the procedure in the N2O/O2 group compared to the sedation and unsedated groups. Forbes 2000 reported that the endoscopists perception of the patients tolerance to the procedure was better in the sedation group.

There was no difference between the two groups with regards to depth of sedation (Maslekar 2009 CD), however, patients experienced more sedation in the sedation group when compared to N2O/O2 and placebo groups (Saunders 1994).

There was also significantly less nausea scores in the N2O/O2 group compared to the sedation group.

None of the studies reported complications due to the coloscopy, however, 2 studies reported significantly higher side effects of the sedation groups compared to the N2O/O2 groups (Notini‐Gudmarsson 1996 and Saunders 1994), while one study showed that the N2O/O2 group had more side effects than the sedation group (Forbes 2000).

Discussion

Though it was not possible to pool the results of each of the study due to the reasons already mentioned, we can deduct from the results of the studies included that nitrous oxide is at least as safe and as effective as different types of sedation and is at least as acceptable to the patients as sedation. Furthermore, two studies report less side‐effects in the N2O/O2 (Notini‐Gudmarsson 1996 and Saunders 1994) and one reported more side effects than the sedation groups (Forbes 2000).

With colonoscopies being one of the most frequently performed procedures today (Maslekar 2009), controlling pain/discomfort and reducing the time required to complete each procedure is of vital importance. Patient safety, procedure efficiency and cost effectiveness are the main stay of most colonoscopy units world wide.

By decreasing the patients' pain and/or discomfort during any procedure the practitioner can complete the procedure with increased precision, less risk of complications and in less time. Though there was no overall majority consensus between the studies with regards to pain/discomfort control, four studies did however show that nitrous oxide was at least as good as various sedation methods in controlling patients symptoms of pain/discomfort (Maslekar 2009 CD, Notini‐Gudmarsson 1996, Lindblom 1994, Saunders 1994).

It was evident through six studies that the patients have a quicker recovery period with nitrous oxide, which therefore leads to less time spent in the assessment unit, which in turn lead to cost reduction, unlike sedation. With one study showing that nitrous oxide patients needed the same amount of time to recover. None of the studies showed that patients receiving nitrous were slower to recover. With the increasing demand of colonoscopy procedures improving cost effectiveness and number of procedures done improved the efficiency of colonoscopy units to meet target demands, in addition to reducing the amount of time patients have to take off of work (Martin 2000).

However, a conclusive decision cannot be taken, a systematic review is only as good as the studies it reviews, collectively these studies are riddled with various biases.

Authors' conclusions

Implications for practice.

The evidence suggests that nitrous oxide does have a role in colonoscopy procedures in controlling pain/discomfort with low risk of adverse effects, quicker recovery and good patients overall satisfaction.

Implications for research.

Further research trials which are all party‐blinded, appropriately randomised, with adequate allocation and concealment, and free of selective reporting or any other source of bias are required. Studies should be conducted in comparing between nitrous oxide with various commonly used sedation methods.

Outcome measures should focus on overall symptom control during and after the procedure and specifically looking at pain/discomfort control. In addition to other parameters such as completion rate, recovery time, cost differences, patients satisfaction, and post discharge follow up.

Only through meticulously conducted trials can a conclusive decision be made as to whether or not nitrous oxide should replace sedation during colonoscopies.

Appendices

Appendix 1. Search strategy for Clib 22.10.2009

#1(Entonox OR Nitronox) or (nitr* ADJ oxide*) or (N2O adj O2) or "N2O/O2" or "gas and air" 194 hits

#2MeSH descriptor Colonoscopy explode all trees 1053 hits

#3(colonoscop* OR sigmoidoscop*) 1779 hits

#4(#2 OR #3) 1779 hits

#5MeSH descriptor Nitrous Oxide explode all trees 1155 hits

#6(#1 OR #5) 1286 hits

#7(#4 AND #6) 15 hits

Appendix 2. MEDLINE 20.10.09

1. randomised controlled trial.pt.

2. controlled clinical trial.pt.

3. randomized.ab.

4. placebo.ab.

5. clinical trial.sh.

6. randomly.ab.

7. trial.ti.

8. 1 or 2 or 3 or 4 or 5 or 6 or 7

9. humans.sh.

10. 8 and 9

11. exp Endoscopy, Gastrointestinal/

12. (endoscop* or colonoscop* or sigmoidoscop*).mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier]

13. 11 or 12

14. exp Nitrous Oxide/

15. (nitr* adj oxide*).mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier]

16. (entonox or nitronox).mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier]

17. ((N2O adj O2) or "N2O/O2").mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier]

18. "gas and air".mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier]

19. 18 or 16 or 17 or 14 or 15

20. 10 and 13 and 19

Appendix 3. Search strategy EMBASE 22.10.09

1. randomised controlled trial/

2. randomizations/

3. controlled study/

4. multicenter study/

5. phase 3 clinical trial/

6. phase 4 clinical trial/

7. double blind procedure/

8. single blind procedure/

9. ((singl* or doubl* or trebl* or tripl*) adj (blind* or mask*)).ti,ab.

10. (random* or cross* over* or factorial* or placebo* or volunteer*).ti,ab.

11. 6 or 3 or 7 or 9 or 2 or 8 or 4 or 1 or 10 or 5

12. "human*".ti,ab.

13. (animal* or nonhuman*).ti,ab.

14. 13 and 12

15. 13 not 14

16. 11 not 15

17. exp gastrointestinal endoscopy/

18. (endoscop* or colonoscop* or sigmoidoscop*).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer name]

19. 17 or 18

20. exp Nitrous Oxide/

21. (nitr* adj oxide*).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer name]

22. (entonox or nitronox).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer name]

23. ((N2O adj O2) or "N2O/O2").mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer name]

24. "gas and air".mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer name]

25. 24 or 22 or 23 or 20 or 21

26. 25 and 16 and 19

Appendix 4. Search Strategy PubMed/Internet 07.01.10

((("Colonoscopy"[Mesh] OR "Endoscopy"[Mesh]) AND "Randomized Controlled Trial "[Publication Type]) AND "Entonox "[Substance Name])

((("Colonoscopy"[Mesh] OR "Endoscopy"[Mesh]) AND "Randomized Controlled Trial "[Publication Type]) AND "Nitrous Oxide"[Mesh])

(("Colonoscopy"[Mesh] AND "Randomized Controlled Trial "[Publication Type]) AND  "Nitrous Oxide"[Mesh])

Data and analyses

Comparison 1. Caecal Intubation Rate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Intubation Rate	7	507	Odds Ratio (M‐H, Fixed, 95% CI)	0.58 [0.25, 1.33]

1.1. Analysis.

Comparison 1 Caecal Intubation Rate, Outcome 1 Intubation Rate.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Forbes 2000.

Methods	Randomized, controlled trial
Participants	102 patients (55 males and 47 females) were included. All patients undergoing elective colonoscopy were included. Patients with a history of colonic resection, chronic respiratory disease, medical problems precluding the self‐administration of Entonox, patients who were non English speaking, or patients undergoing combined upper GI endoscopy and colonoscopy were excluded.
Interventions	Patients received either intravenous midazolam (0.06mg/kg) and meperidine (0.76mg/kg) or Entonox (inhaled through a mouth piece for 2 minutes initially and then continually until the caecum was reached, then as required). Further doses of sedation were given if required in the sedation group. In the N2O group, if the patients was too uncomfortable or the colonoscopists judged that the patients was in discomfort intravenous midazolam and meperidine was given after a 5 minute washout period.
Outcomes	All patients completed an Anxiety on a 100mm visual analogue scale (VAS). VAS measurements were taken before discharge on the following: ‐Pain ‐Tolerance ‐Satisfaction ‐would you have a colonoscopy again Patients were also assessed on whether they felt nausea, dizziness, drowsiness, headache, abdominal discomfort, and tingling or numbness in the limbs. Other parameters measure: ‐ colonoscopists' satisfaction (using VAS) ‐ colonoscopists' assessment of procedure difficulty ‐ time taken to reach the caecum ‐ completion rate ‐ time to discharge (from time completed colonoscopy to time considered fit for discharge); this was decided by use of memory card test and letter deletion test.
Notes	No patients dropped out. 20 patients developed adverse effects during the colonoscopy (15 in the Entonox group (2 hypotension, 9 O2 desaturation, 5 marked sweating, 1 nausea, 1 hallucinations) and 5 in the sedation group (3 hypotension and 2 marked sweating). 3 patients developed severe adverse effects after the colonoscopy (2 in the sedation group: severe drowsiness and severe abdominal discomfort and 1 in the Entonox group: severe abdominal discomfort.)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No mention of generation of allocation sequence.
Allocation concealment (selection bias)	Unclear risk	No mention of allocation concealment.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	No mention of blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No Missing outcome data.
Selective reporting (reporting bias)	Low risk	There were no indications of selective reporting.
Other bias	High risk	The unclarity of the blinding, sequence generation and allocation concealment is considered a source of bias. The endoscopists opinion of the patients tolerability of the procedure can be considered a source of bias.

Lindblom 1994.

Methods	Randomized, placebo controlled, double blinded trial
Participants	50 patients (26 males and 24 females) were included. All patients undergoing elective colonoscopy were included. Exclusion criteria: previous colonic surgery, drug abuse, intolerance to any of the drugs in the study, or inability to speak Swedish.
Interventions	Patients received either intravenous ketobemidone hydrochloride (2.5mg) and midazolam (2.5mg) in addition to inhaled O2 or intravenous saline and inhaled N2O/O2 (patient controlled Entonox). If the patients in either groups got too uncomfortable or felt pain, intravenous ketobemidone hydrochloride.
Outcomes	Patients were asked to rank their discomfort during the colonoscopy on a VAS on leaving the unit (0=no discomfort, 100=worst discomfort imaginable). The total colonoscopy time, the colonoscopy completion rate, complications and the time for recovery were also measured. A modified recollection test (memory test) was performed by the research nurse before the colonoscopy and at the time the patient is leaving the unit to assess procedure related and medication related effects.
Notes	No complications occurred.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomizations was used; block size of ten.
Allocation concealment (selection bias)	Low risk	Closed envelopes.
Blinding (performance bias and detection bias) All outcomes	Low risk	‘Double‐dummy’ blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	No Missing outcome data.
Selective reporting (reporting bias)	Low risk	There were no indications of selective reporting.
Other bias	Low risk	There were no other sources of bias identified.

Maslekar 2009 BJS.

Methods	Randomized, controlled trial
Participants	131 patients (63 males and 68 females) were included. All patients undergoing elective colonoscopy were included. Patients with a history of pneumothorax, severe asthma, previous colonic resection, intolerance to any of the drugs, those undergoing upper GI endoscopy and colonoscopy simultaneously, and those taking opiates were excluded.
Interventions	Patients received either intravenous midazolam (2mg/ml max 10mg) and fentanyl (0.075‐0.1mg) or N2O/O2 (Entonox) inhaled through a mouth piece for 2 minutes initially and then continually until the caecum was reached, then as required. Further doses of midazolam were given if required in the sedation group. In the N2O group, if the patients was too uncomfortable or the colonoscopists judged that the patients was in discomfort intravenous sedation was given after a 5 minute washout period. 3 types of colonoscopists performed the procedure; consultants, senior nurses (>5000 colonoscopy experience), and non‐medical colonoscopists.
Outcomes	All patients completed an Anxiety and Depression questionnaire, baseline letter cancellation test (which measure concentration and perception), and scored pain on a 100mm visual analogue scale (VAS). VAS measurements of pain experience were taken immediately and at 15 minute intervals after the procedure until discharge. Patients also marked VAS at 24 hours after the procedure to allow for recovery. Other parameters measured: ‐ degree of sedation (measured on the Modified Observer's Assessment of Alertness/Sedation Scale (MOAAS) at 1 minute intervals during the procedure and at 5 minute intervals during recovery) ‐ patient, nurse, and colonoscopist satisfaction (using VAS) ‐ complication rate ‐ colonoscopist's assessment of procedure difficulty ‐ time taken to reach the caecum and total colonoscopy time ‐ completion rate (Completion of the colonoscopy was based on 2 of 3 landmarks: the ileocaecal valve, the appendiceal orifice, and/or the triradiate fold.) ‐ degree of psychomotor recovery (assessed by using the letter cancellation test immediately upon arrival to the recovery area and then at 15 minute intervals until discharge) ‐ time to discharge
Notes	No patients dropped out. 1 patients in the sedation group developed Nausea but resolved spontaneously.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomizations was used; block size of six.
Allocation concealment (selection bias)	Low risk	Sequentially numbered, opaque sealed envelopes.
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding was done
Incomplete outcome data (attrition bias) All outcomes	Low risk	No Missing outcome data.
Selective reporting (reporting bias)	Low risk	There were no indications of selective reporting.
Other bias	High risk	No blinding is a source of bias.

Maslekar 2009 CD.

Methods	Randomized, controlled trial
Participants	100 patients (53 males and 47 females) were included. All patients undergoing elective colonoscopy were included. Patients with a chronic pulmonary disease, previous colonic resection, intolerance to any of the drugs, unwilling to enter the trial, ASA class IV, allergy to soybeans or eggs, history of seizure disorder, sleep apnoea, difficult intubation, short thick neck, and inability to open mouth widely were excluded.
Interventions	Patients received either intravenous Propofol (through a pump to achieve target concentration of 1.2 µg/ml/hr; the procedure would be started after reaching a sedation (MOAAS) score of 4) in addition to a patient controlled analgesia pump containing Propofol (which delivered a bolus of 200mcg/kg/ml, with a lockout period of 2 minutes) or N2O/O2 (Entonox) inhaled through a mouth piece; the procedure would be started after reaching a sedation (MOAAS) score of 4, the patient continued to inhale Entonox until the caecum was reached, then as required. Intravenous Fentanyl were given if required in the sedation group. In the N2O group, if the patients was too uncomfortable or the colonoscopists judged that the patients was in discomfort intravenous sedation was given after a 5 minute washout period.
Outcomes	All patients completed an Anxiety and Depression questionnaire, baseline letter cancellation test (which measure concentration and perception), and scored pain on a 100mm visual analogue scale (VAS). VAS measurements of pain experience were taken immediately and at 15 minute intervals after the procedure until discharge. Patients also marked VAS at 24 hours after the procedure to allow for recovery. Other parameters measure: ‐ degree of sedation (measured on the Modified Observer's Assessment of Alertness/Sedation Scale (MOAAS) at 1 minute intervals during the procedure and at 5 minute intervals during recovery) ‐ patient, nurse, and colonoscopist satisfaction (using VAS) ‐ complication rate (Sedation complication: prolonged drop in O2 saturation below 90% for at least 30 seconds, with the need for positive pressure ventilation using a bag‐valve system. Others: prolonged drop in blood pressure below 90mm Hg, Heart rate below 60 or more than 110 beats/min.) ‐ colonoscopist's assessment of procedure difficulty ‐ time taken to reach the caecum and total colonoscopy time ‐ completion rate (Completion of the colonoscopy was based on 2 of 3 landmarks: the ileocaecal valve, the appendiceal orifice, and/or the triradiate fold.) ‐ degree of psychomotor recovery (assessed by using the letter cancellation test immediately upon arrival to the recovery area and then at 15 minute intervals until discharge) ‐ time to discharge
Notes	No patients dropped out. No mention of whether or not complications occurred.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomizations was used; block size of five.
Allocation concealment (selection bias)	Low risk	Sequentially numbered, opaque sealed envelopes.
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding was done
Incomplete outcome data (attrition bias) All outcomes	Low risk	No Missing outcome data.
Selective reporting (reporting bias)	Low risk	There were no indications of selective reporting.
Other bias	High risk	No blinding is a source of bias.

Notini‐Gudmarsson 1996.

Methods	Randomized, controlled trial
Participants	38 patients were included. All patients undergoing elective colonoscopy were included.
Interventions	Patients received either intramuscular pethidine hydrochloride (1mg/kg) 20‐30 min before the procedure or Medimix (N2O/O2 mixture) (inhaled through a mouth piece just before the procedure and then continually until the caecum was reached, then as required). Atropine (0.5mg IM) was given to all patients 30minutes before the procedure. If the patients in either groups got too uncomfortable or felt pain, intravenous midazolam (2.5mg).
Outcomes	The colonoscopist assessed three parameters: technical conditions (easier, as expected, more difficult), time consumption (less, as expected, more time), and instrument insertion depth (less, as expected, more proximal), all three scores were added to rank the overall colonoscopists’ assessment. The anaesthetist measure blood pressure, pulse rates, and O2 saturation immediately before, during and five minutes after the colonoscopy. The completion time (from the start of the procedure until discharge) was documented. (discharge criteria: correctly performed co‐ordination test‐standing and touching the nose with closed eyes‐and absence of any unpleasant sensation) The patients filled in a questionnaire at the end of colonoscopy and at discharge, rating on a VAS any nausea or pain felt (0=none; 10 worst ever) in addition to a general evaluation (0=good experience; 100=worst possible experience). Other parameters ranked were: dizziness, fatigue, and headaches.
Notes	2 patients required IV fluids due to dehydration caused by the laxative bowel preparation.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No mention of how the allocation sequence was generated.
Allocation concealment (selection bias)	Unclear risk	No mention of allocation concealment.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	No mention of blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No Missing outcome data.
Selective reporting (reporting bias)	Low risk	There were no indications of selective reporting.
Other bias	High risk	The unclarity of the blinding, sequence generation and allocation concealment is considered a source of bias.

Saunders 1994.

Methods	Randomized, placebo controlled, double blind trial
Participants	89 patients (38 males and 51 females) were included. All patients undergoing elective colonoscopy were included. Patients with a history of pneumothorax, chronic obstructive airways disease, bowel obstruction, or malignant hyperpyrexia were excluded.
Interventions	Patients received either intravenous midazolam (2.5mg), pethidine (50mg), and hyoscine butylbromide (20mg), plus inhaled air or N2O/O2 mixture (hyoscine butylbromide 20mg IV plus Entonox) or placebo (hyoscine butylbromide 20mg IV plus inhaled air). Patients were instructed to breathe the allocated gas continuously for 1 min. Additional breathes of gas were taken if discomfort occurred. Further doses of midazolam (1‐2.5mg) and pethidine (25‐50mg) were given if the patients was too uncomfortable.
Outcomes	All patients marked their degree of anxiety on a 100mm visual analogue scale (0=totally calm; 100=extremely anxious) before the procedure, in addition to base line recording of pulse blood pressure, and O2 saturation. An independent observer recorded episodes of patient discomfort, blood pressure (at 5min intervals) O2 saturation, and the time and extent of the intubation. After each procedure the colonoscopists recorded the difficulty of the procedure (easy, average, difficult). Patients were discharged by assess physiological parameters, verbal and motor responses. The total time to discharge was also documented. A VAS was used to assess patients level of pain (0=no pain; 100=agony) and sedation (0=wide awake; 100 asleep). Patients were also asked to complete a questionnaire regarding their experience and specifically asked about nausea, vomiting, paraesthesia, dizziness, drowsiness, and headaches.
Notes	Unable to extract data from charts, no mention of ranges or SE.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomizations was used; block size of six.
Allocation concealment (selection bias)	Unclear risk	Authors mention that the randomizations was stratified, but no mention on how.
Blinding (performance bias and detection bias) All outcomes	Low risk	All endoscopy personnel were blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	No Missing outcome data.
Selective reporting (reporting bias)	Low risk	There were no indications of selective reporting.
Other bias	High risk	The unclarity of the allocation concealment is considered a source of bias.

Trojan 1997.

Methods	Randomized, controlled trial
Participants	27 patients (17 males and 10 females) were included.
Interventions	Patients received either intravenous midazolam (2.5mg), pethidine (25‐50mg), and hyoscine butylbromide (20mg as premedication) or N2O/O2 mixture (hyoscine butylbromide 20mg(as premedication) IV plus Entonox). 10 patients were also included to receive only hyoscine butylbromide (20mg), but were not included in the randomizations. Patients were instructed to breathe the allocated gas continuously for 1 min. Additional breathes of gas were taken if needed.
Outcomes	After each procedure the colonoscopists recorded the difficulty of the procedure (easy, average, difficult) and toleration of the procedure (good, moderate, or poor), as well as the extent of intubation and the total required examination time. A series of psychomotor tests were carried out, 20 minutes before the colonoscopy, immediately after the procedure and then at 15, 30, 45 minutes after the colonoscopy. The tests were tracking test (for psychomotor co‐ordination skills), multiple choice reaction time test (for reaction time), manual dexterity test (for fine motor control), and letter cancellation test (for perception).
Notes	Only data of the patients that were randomised was collected. Unable to extract data from charts and no mention of individual results in the results section.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No mention of how the allocation sequence was generated.
Allocation concealment (selection bias)	Unclear risk	No mention of allocation concealment.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	No mention if the colonoscopists or the patients were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Incomplete data was addressed.
Selective reporting (reporting bias)	Low risk	There were no indications of selective reporting.
Other bias	High risk	The unclarity of the blinding, sequence generation and allocation concealment is considered a source of bias. The endoscopists opinion of the patients tolerability of the procedure can be considered a source of bias.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Eckardt 1999	Not a RCT; Study did not use N2O.
Fink 1998	Not a RCT; Comment on Trojan 1997
Maslekar 2006	Not a RCT; abstract
Maslekar 2007	Not a RCT; abstract
Michaud 1999	Not a RCT; study on patients who were undergoing upper GI endoscopy or recto‐sigmoidoscopy.
Phillips 1995	Not a RCT; descriptive article of different drugs that are used during colonoscopy procedures.
Ristikankare 1999	A randomised trial, but does not look at N2O; compares sedation with midazolam, to sedation with saline to no sedation.
Terruzzi 2001	RCT comparing regular sedation and analgesia versus on demand sedation and analgesia, but N2O was not looked at as one of the comparative measures.
Theodorou 2001	RCT, but compared 2 different anaesthetic techniques for ambulatory colonoscopy. 1st group received intravenous fentanyl, midazolam, and propofol combination and 2nd group received sevoflurane (which is an anaesthetic agent) in 67% N2O. As the study did not compare N2O/O2 mixture without adding an anaesthetic against the study was excluded.
Trojan 1998	Not a RCT; abstract
Trojan 1998GE	Not a RCT; abstract

Contributions of authors

‐ see above

Sources of support

Internal sources

• Cochrane Colorectal Cancer Group, Denmark.

  Support with search strategy.

External sources

• No sources of support supplied

Declarations of interest

None of the authors have any conflict of interest known.

New