Oxygenation during the apnoeic phase preceding intubation in adults in pre‐hospital, emergency department, intensive care and operating theatre environments

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

To assess the effects of apnoeic oxygenation before intubation in adults in the operating theatre, emergency department, intensive care unit (ICU), or pre‐hospital settings.

Primary objective

To assess the effect of oxygenation compared with no oxygenation during the apnoeic phase of intubation on length of hospital stay and peri‐intubation critical desaturation, in adults in pre‐hospital, emergency department, intensive care and operating theatre environments. Critical desaturation will be deemed to be a reduction in saturation of pulse oximetry (SpO₂) to less than 80%.

Secondary objectives

To assess the effect of oxygenation compared with no oxygenation during the apnoeic phase of intubation on the incidence of desaturation (SpO₂ below 93%), lowest recorded SpO₂, ICU length of stay, first‐attempt intubation success rate and adverse event rate during intubation (e.g. cardiac arrhythmia), in adults in pre‐hospital, emergency department, intensive care and operating theatre environments.

Background

Description of the condition

Endotracheal intubation is an essential procedure for the establishment of a definitive airway in a number of clinical settings, including elective surgery, emergency departments, intensive care units or pre‐hospital settings. The Royal College of Anaesthetists and Difficult Airway Society's Fourth National Audit Project (NAP4) suggests that endotracheal intubation is used in 38.4% of general anaesthetics in the UK (Woodall 2011). Endotracheal intubation involves the placement of a cuffed endotracheal tube into the trachea, typically through the mouth, to enable oxygenation, ventilation and prevention of aspiration. Prolonged attempts at endotracheal intubation are associated with desaturation, unplanned admission to the intensive care unit and death (Caplan 1990; King 1990; Rose 1994). Hypoxaemia in this context is a potentially catastrophic complication and can lead to dysrhythmia, hypoxic brain injury and death (Caplan 1990; Davis 2008; Mort 2004).

Preoxygenation is a widely accepted technique for prolonging the safe apnoea time (Weingart 2012), i.e. the period during the intubation procedure, during which the patient remains apnoeic without desaturation. Preoxygenation can be less effective in critically ill patients due to the effects of shunt physiology, increased metabolic demand, anaemia and decreased cardiac output (Drummond 1984; Farmery 1996; Mort 2005). In the NAP4, nearly 20% of all airway incidents occurred in the intensive care unit, and ineffective preoxygenation in the critically ill was suggested as one possible contributor to this (Woodall 2011).

Description of the intervention

Apnoeic oxygenation aims to prevent desaturation via the application of various oxygen delivery techniques during the apnoeic phase of intubation (Weingart 2012). Various protocols have been employed with nasal cannulae, including humidified or non‐humidified oxygen, as well as a range of oxygen flow rates (Binks 2017a; Binks 2017b; Holyoak 2017; White 2017). Low‐flow nasal cannulae (LFNC) are traditionally used in patients requiring minimal oxygen supplementation. These are nasal cannulae attached to wall oxygen, traditionally titrated at flows of 1 to 6 litres per minute of dry oxygen in spontaneously breathing patients (O'Driscoll 2008). Oxygen can be delivered via LFNC at up to 15 litres per minute in the setting of apnoeic oxygenation (Binks 2017a). High‐flow nasal cannulae (HFNC) are used as a non‐invasive delivery system for humidified oxygen and are capable of delivering oxygen flow rates of up to 60 litres per minute or more (Corley 2017; Mir 2017).

How the intervention might work

Alveolar capillary blood continues to absorb oxygen from the alveoli in the absence of lung movement, leading to a net subatmospheric alveolar pressure (Weingart 2012). During apnoeic oxygenation, the continuous oxygen supply is theorized to continue into the alveoli via a mass flow effect (Weingart 2012). Additionally, the use of apnoeic oxygenation, particularly via HFNC, can produce a positive distension pressure within the bronchioles and alveoli (Humphreys 2017). Therefore apnoeic oxygenation may reduce the extent of atelectasis on induction of anaesthesia (Humphreys 2017).

Why it is important to do this review

Respiratory complications of endotracheal intubation have the potential to lead to significant adverse events including dysrhythmia, haemodynamic decompensation, hypoxic brain injury and death (Davis 2008; Mort 2004). Apnoeic oxygenation may serve as a non‐invasive adjunct to endotracheal intubation to decrease the incidence of desaturation, morbidity and mortality.

Objectives

To assess the effects of apnoeic oxygenation before intubation in adults in the operating theatre, emergency department, intensive care unit (ICU), or pre‐hospital settings.

Primary objective

To assess the effect of oxygenation compared with no oxygenation during the apnoeic phase of intubation on length of hospital stay and peri‐intubation critical desaturation, in adults in pre‐hospital, emergency department, intensive care and operating theatre environments. Critical desaturation will be deemed to be a reduction in saturation of pulse oximetry (SpO₂) to less than 80%.

Secondary objectives

To assess the effect of oxygenation compared with no oxygenation during the apnoeic phase of intubation on the incidence of desaturation (SpO₂ below 93%), lowest recorded SpO₂, ICU length of stay, first‐attempt intubation success rate and adverse event rate during intubation (e.g. cardiac arrhythmia), in adults in pre‐hospital, emergency department, intensive care and operating theatre environments.

Methods

Criteria for considering studies for this review

Types of studies

We will include all randomized controlled trials (RCTs) and quasi‐RCTs that compare apnoeic oxygenation to no apnoeic oxygenation for endotracheal intubation. We define quasi‐randomization as non‐true randomization to each arm by means such as alternation, case record number or date of birth. We will exclude all comparative prospective cohort and comparative retrospective cohort studies. We will exclude physiological modelling studies or case reports. We will not impose restrictions on language or publication status.

Types of participants

We will include trials involving adults aged 18 years and older who require intubation in the operating theatre, ICU, emergency department (ED) and pre‐hospital environments.

Types of interventions

We will include studies that compare the use of any form of apnoeic oxygenation — including HFNC and LFNC — versus no apnoeic oxygenation during intubation.

Types of outcome measures

Critical outcomes

1. Hospital stay (days)

2. Incidence of severe hypoxaemia (SpO₂ below 80%) from start of apnoea until successful intubation

Important outcomes

1. Incidence of hypoxaemia (SpO₂ below 93%) from start of apnoea until successful intubation

2. Lowest recorded SpO₂ between the start of intubation and recovery from anaesthesia (SpO₂%)

3. Intensive care unit stay (days)

4. First pass success rate (proportion of successful intubation on first laryngoscopy attempt)

5. Adverse events, including:

   1. pharyngeal or dental trauma during intubation;

   2. cardiac arrhythmias and cardiac arrest during intubation;

   3. aspiration pneumonia (as defined by the study authors) within 24 hours of procedure;

   4. other serious adverse events reported by study authors.

6. Mortality (in‐hospital or 30‐day, as defined by study authors)

We will screen all included studies to identify relevant post hoc outcomes.

Search methods for identification of studies

Electronic searches

We will search electronic databases for articles using key words, synonyms and subject headings that relate to apnoeic oxygenation and intubation. We will use controlled vocabulary specific to each database. We will use the Cochrane Highly Sensitive Search Strategy for the identification of RCTs (Lefebvre 2011). Two review authors (CT and RV) will perform the search independently and no language restrictions will be applied. We will contact individual trial authors for additional information where necessary.

We will search the following databases for published trials:

1. Cochrane Central Register of Controlled Trials (CENTRAL)(1996 to present);

2. MEDLINE Ovid (1946 to present) (Appendix 1);

3. Elsevier Embase (1947 to present);

4. Institute for Scientific Information Web of Science (1900 to present);

5. CINAHL EBSCO (1937 to present).

We will also search the following major clinical trial registers:

1. World Health Organization International Clinical Trials Registry Platform (ICTRP) apps.who.int/trialsearch/);

2. ClinicalTrials.gov (www.clinicaltrials.gov/);

3. ISRCTN registry (www.isrctn.com/).

Searching other resources

We will perform a citation search of all included studies as well as any relevant studies and reviews on apnoeic oxygenation. If this identifies any additional eligible studies we will re‐examine and update the search strategy. We will also perform a search of Open Grey for grey literature, and Google Scholar for additional trials. We will consider for inclusion studies that are available as abstracts only.

Data collection and analysis

We will use standard methods and follow Cochrane guidelines, as outlined in the following sections.

Selection of studies

Two review authors (CT and RV) will independently screen all titles and abstracts of each reference identified by our search, and will then independently assess the full text of any potentially relevant studies for eligibility or exclusion. We will use Covidence software to collate search results, remove duplicates and record screening decisions and exclusions at each stage. This will be facilitated by a standardized digital screening, inclusion and exclusion tools, in line with methods outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019a). We will resolve any disagreements by discussion and consensus prior to proceeding at each stage.

We will record the process in sufficient detail to produce a PRISMA flow diagram (Moher 2009). For all full‐text articles screened and excluded, we will record the reason for exclusion in sufficient detail to present a table of characteristics for excluded studies.

Data extraction and management

Through Covidence, we will use a standardized digital data‐extraction sheet, in line with requirements outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019a). Two review authors (LW and TM) will independently extract information regarding trial design, comparator, intervention, assessment of methodology, data for comparison, and any additional relevant information. Where required, we (LW and TM) will contact individual trial authors or organizations to obtain missing data or clarification regarding unclear data. We will resolve any disagreements by discussion and consensus.

Assessment of risk of bias in included studies

Two review authors (LW and TM) will independently assess methodological quality. For all included trials we will utilize Cochrane's tool for assessing the risk of bias for RCTs (Higgins 2019a). There are seven specific measurements covered by this tool, and the Cochrane Handbook for Systematic Reviews of Interventions describes how each domain should be assessed to reach a judgement of low, high or unclear risk of bias. If a study is judged to be low risk in all domains then it will be deemed to be at low risk of bias overall. The criteria include:

1. random sequence generation;

2. allocation concealment;

3. blinding of participants and personnel;

4. blinding of outcome assessment;

5. incomplete outcome data;

6. selective reporting;

7. other sources of bias.

We will resolve any disagreements by discussion and consensus. To assess selective reporting, we will screen the clinical trial databases ISRCTN (www.isrctn.com) and ClinicalTrials.gov for protocols of all included studies. We will present a 'Risk of bias' summary, as well as 'Risk of bias' judgements for individual studies in the 'Characteristics of included studies' tables.

Measures of treatment effect

For dichotomous outcomes, we will calculate risk ratios (RRs) with 95% confidence intervals (CIs). For continuous outcomes, we will calculate mean differences (MDs) with 95% CIs. When appropriate, we plan to calculate a number needed to treat for an additional beneficial outcome (NNTB) or a number needed to treat for an additional harmful outcome (NNTH) to aid clinical decision making.

Unit of analysis issues

The standard unit of analysis for our outcomes will be the participant. In the event of multiple intervention studies, we will conduct pair‐wise comparisons (A versus control; B versus control, etc.). We will not include a group twice in the same analysis. If multiple interventions are to be included in the same analysis, we will combine them into a single intervention group.

If we identify a cluster‐randomized or cross‐over trial as eligible for inclusion, then we will decide specifically on whether that trial can be included in the statistical analysis, based on the availability of specific information either in published material or from correspondence with the trial authors. If there are multiple time points reported in a single study, we will utilize the most relevant for this review (usually the pre‐intervention baseline and first time point post‐intervention). Our decisions will be made in line with the recommendations in Chapter 23 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019b).

Dealing with missing data

We will contact trial authors to request missing data. We will outline any missing data in the results and explain in the discussion the extent to which the overall results may be affected. In the event of unreported standard deviations, we will make an estimate based on other reports within the meta‐analysis.

Assessment of heterogeneity

If there is obvious extensive clinical heterogeneity, we will not perform a meta‐analysis. Otherwise, we will investigate clinical heterogeneity via meta‐regression, identifying trends within subgroups of the clinical population.

We will utilize the I² test to assess statistical heterogeneity for each outcome (Higgins 2019a). We will judge I² values as follows;

• 0% to 40%: might not be important;

• 30% to 60%: may represent moderate heterogeneity;

• 50% to 90%: may represent substantial heterogeneity;

• 75% to 100%: considerable heterogeneity.

We will explore any outcomes associated with substantial or considerable heterogeneity for potential contributing factors.

Assessment of reporting biases

Where possible, if there are more than 10 included studies, we will perform a visual assessment of a funnel plot to assess reporting bias. This will be done according to the guidance in theCochrane Handbook for Systematic Reviews of Interventions (Higgins 2019a).

We will attempt to minimise the impact of reporting bias, and assess its potential impact, by searching clinical trial databases (i.e. www.isrctn.com and ClinicalTrials.gov) for the results of registered trials that are completed but not yet published.

Data synthesis

We will analyse the data using Review Manager 5 (Review Manager 2014). We will apply the fixed‐effect Mantel‐Haenszel model to all analyses unless there is significant heterogeneity (I² P value below 0.05). If this occurs then we will use the inverse variance method to apply a random‐effects model. We will assume statistical significance to correspond to an alpha of less than 0.05. For our outcome measures, we will present data in forest plots and will report 95% CIs.

Subgroup analysis and investigation of heterogeneity

Where sufficient data permit for each outcome, we will perform subgroup analyses by meta‐regression to investigate heterogeneity. In particular, we plan to explore the effect of location of intubation (ED, ICU, operating theatre or pre‐hospital), since this is likely to be a cause of heterogeneity. Indications for intubation and the skill of the clinician will differ significantly between each of these settings (Binks 2017a; Holyoak 2017).

We also plan to assess whether there is a significant effect of flow rate, as a categorical — i.e. low flow (less than 15 L/min) versus high flow (more than 15 L/min) — and continuous variable. The ability for oxygen to reach the alveolus is likely to be altered by flow rates and may impact the efficacy of apnoeic oxygenation (Holyoak 2017).

Lastly, we intend to assess whether there are particular patient factors which may affect this outcome (e.g. obesity or pregnancy). Obstetric and obese patients have significant alterations in their respiratory anatomy and physiology that are likely to alter the efficacy of apnoeic oxygenation (McClelland 2009; Ramachandran 2009).

These factors will be meta‐regressed in a random‐effects model using Stata (Stata 2011). This model will use a Knapp‐Hartung modification with residual restricted maximum likelihood to measure between‐study variance (τ2) (Higgins 2004).

Sensitivity analysis

We will perform a sensitivity analysis of studies to explore the influence of 'Risk of bias' assessments (see Assessment of risk of bias in included studies). This will allow assessment of the robustness of our conclusions. These sensitivity analyses will be limited to the primary outcomes.

Summary of findings and assessment of the certainty of the evidence

The eight GRADE considerations (study limitations, consistency of effect, imprecision, indirectness, publication bias, magnitude of effect, possible confounders and dose‐response gradient) will be used to assess the certainty of the body of evidence presented in this review (Guyatt 2008). The outcomes assessed using GRADE will include hospital length of stay, incidence of severe hypoxaemia, ICU length of stay, and first‐pass intubation success. This will be done in accordance with guidance in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019a). We will use GRADEpro GDT software to perform this task (GRADEpro GDT). All decisions to upgrade or downgrade our assessment of the certainty of the evidence will be described clearly as footnotes to the table.

Appendices

Appendix 1. Ovid MEDLINE search strategy

1 exp oxygen/
 2 oxygen inhalation therapy/
 3 apnea/
 4 (1 or 2) and 3
 5 (apn?e* adj5 (oxygen* or ventilat*)).af.
 6 Transnasal Humidified Rapid‐Insufflation Ventilatory Exchange.af.
 7 THRIVE.af.
 8 optiflow.af.
 9 ((high flow or high frequency) adj3 (nasal or nasopharyng*)).af.
 10 ((nasal or nasopharyngeal or intranasal or transnasal) adj3 (cannula* or prong* or oxygen*)).af.
 11 (hfnc or hfnp or hhfnox or hfno).af.
 12 (preoxygenat* or pre oxygenat*).af.
 13 (passive adj2 oxygen*).af.
 14 (diffusion adj2 respirat*).af.
 15 (oxygen* adj2 insufflat*).af.
 16 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15
 17 exp intubation, intratracheal/
 18 laryngoscopy/
 19 exp respiration artificial/
 20 exp anesthesia, general/
 21 ((noninvasive or non invasive or artificial or mechanical) adj2 (respirat* or ventilat*)).af.
 22 (nppv or ninppv or nippv).af. (1008)
 23 (Endotracheal or Intratracheal or Intubat* or laryngoscop*).af.
 24 An?esth*.af.
 25 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24
 26 16 and 25
 27 ((randomized controlled trial or controlled clinical trial).pt. or randomized.ab. or placebo.ab. or drug therapy.fs. or randomly.ab. or trial.ab. or groups.ab.) not (animals not (humans and animals)).sh.
 28 26 and 27

Contributions of authors

Authors: Leigh White (LW), Ruan Vlok (RV), Christopher Thang (CT), Mario D'Souza (MD) and Thomas Melhuish (TM).
 Conceiving the review protocol: LW, RV, TM
 Designing the review protocol: LW, RV, CT and TM
 Co‐ordinating the review protocol: LW
 Review Manager statistical data planning: LW, TM and MD
 Other statistical analysis protocol preparation not using Review Manager 5: LW , TM and MD
 Writing the review protocol: all authors
 Providing guidance on the review: nil required
 Securing funding for the review: nil required
 Performing previous work that was the foundation of the present study: LW, RV and TM
 Guarantor for the review (one author): LW
 Person responsible for reading and checking review before submission: LW

Declarations of interest

Leigh D White has no conflicts of interest to declare.

Ruan A Vlok has no conflicts of interest to declare.

Christopher YC Thang has received personal payments indirectly through the University of Queensland in 2018 from consultancy fees for Portal Surgical Procedures, a commercial company. Portal Surgical Procedures is a laparoscopic wound closure device company whose product is not relevant to this review.

Thomas M Melhuish has no conflicts of interest to declare.

Mario D'Souza has no conflicts of interest to declare.

New