Apnoeic oxygenation during paediatric tracheal intubation: a systematic review and meta-analysis

Background: Supplemental oxygen administration by apnoeic oxygenation during laryngoscopy for tracheal intubation is intended to prolong safe apnoea time, reduce the risk of hypoxaemia, and increase the success rate of ﬁrst-attempt tracheal intubation under general anaesthesia. This systematic review examined the efﬁcacy and effectiveness of apnoeic oxygenation during tracheal intubation in children. Methods: This systematic review and meta-analysis included randomised controlled trials and non-randomised studies in paediatric patients requiring tracheal intubation, evaluating apnoeic oxygenation by any method compared with patients without apnoeic oxygenation. Searched databases were MEDLINE, Embase, Cochrane Library, CINAHL, ClinicalTrials.gov, International Clinical Trials Registry Platform (ICTRP), Scopus

Supplemental oxygen administration by apnoeic oxygenation during laryngoscopy for tracheal intubation could prolong safe apnoea time, reduce the risk of hypoxaemia, and increase the success rate of first attempt tracheal intubation under general anaesthesia.This systematic review examined the efficacy of apnoeic oxygenation in facilitating tracheal intubation in children.Apnoeic oxygenation increased intubation first-pass success, increased oxygen saturation during intubation, and decreased incidence of hypoxaemia compared with no supplementary oxygen.However, the included studies were heterogeneous, and more high-quality randomised controlled trials are warranted with well-defined outcome variables.
Tracheal intubation aims to establish a patent airway to ensure ventilation of the lungs during surgery, procedures, or respiratory insufficiency.It is lifesaving for children with severe acute respiratory failure.The number of tracheal intubation attempts is associated with an increased incidence of severe complications. 1,2Interventions to improve firstattempt intubation success rate and increase safety are pivotal. 3Tracheal intubation in the operating theatre is associated with an incidence of difficult airway of ~0.9% in children up to 16 yr old, 4 but alarmingly this is 5.8% in neonates, 1 and is mostly unanticipated.Furthermore, the incidence of haemoglobin desaturation appears to be as high as 40%, potentially leading to severe adverse events in neonates. 1Prolonged severe desaturation can lead to hypoxic encephalopathy, 5 cardiac arrest, 6,7 or death. 8acemask preoxygenation in children can be difficult because of lack of cooperation leading to improper mask seal, 9 and is relatively ineffective, especially in infants, considering hypoxaemia occurs within seconds after cessation of spontaneous or assisted ventilation. 10Safe apnoea time is generally described as the time from cessation of breathing or ventilation until a patient attains a critical level of peripheral arterial oxygen saturation. 11Beyond this point, oxygenation decreases rapidly to critically low blood and tissue oxygen levels endangering vital functions. 12pnoeic oxygenation to reduce the risk of hypoxaemia and extend safe apnoea time was initially described in the early 1900s. 13,14It consists of administering a constant stream of oxygen 100% while the patient is not breathing.The physiological requirements for adequate apnoeic oxygenation are a patent upper and lower airway, diffusion of highly concentrated oxygen into the alveoli, minimal pulmonary shunting, and presence of cardiac function.Low-flow oxygen with different delivery techniques achieves this objective, and mounting evidence supports the possible advantages of highflow nasal oxygen administration. 15However, apnoeic oxygenation is currently not a routine practice during tracheal intubation in most institutions. 2,16,17This systematic review evaluates the efficacy of apnoeic oxygenation on first-attempt intubation success, oxygen saturation, and adverse effects during paediatric tracheal intubation.

Methods
The review protocol for this systematic review and metaanalysis was registered with PROSPERO (registration number: CRD42022369000) on December 2, 2022.We report our findings according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines (PRISMA). 18

Eligibility criteria
We included peer-reviewed randomised controlled trials (RCTs) and non-randomised studies (non-RCTs, interrupted time series, controlled before-and-after studies, and cohort studies) in paediatric patients (age <16 yr) requiring tracheal intubation.Included studies compared apnoeic oxygenation by any method or device with a control group without apnoeic oxygenation.Apnoeic oxygenation was defined as any passive insufflation of oxygen of any flow into the nose or mouth without ventilation.Unpublished studies, case series, conference abstracts, trial protocols, duplicates, and unretrievable articles were excluded.

Searched databases and search strategy
A literature search strategy was devised for the following databases: MEDLINE, Embase, Cochrane Library, CINAHL, Web of Science Core Collection, Scopus, ClinicalTrials.gov,and International Clinical Trials Registry Platform (ICTRP).A medical information specialist (MvG) developed an initial search strategy in MEDLINE with a test against a list of core references to ensure that key publications were included.After refinement, the information specialist set up the search strategy for each information source based on database-specific index terms and free text.The free text search included synonyms, acronyms, and similar terms.No database-provided limits have been applied in any sources considering study types, languages, publication years, or any other formal criteria.The search was finalised on March 22, 2023.The detailed final search strategies were published, 19 and additional information about the systematic search is available in the Supplementary material.

Study selection and assessment
After identifying relevant publications, all were imported into EndNote (EndNote20, Clarivate, Philadelphia, PA, USA).We used Deduklick (Risklick AG, Bern, Switzerland) for deduplication (MvG).An equal number of titles and abstracts were distributed to four groups with two study researchers (ND and ACL, AF and GK, TR and JA, RB and CSR).The researchers independently screened all titles and abstracts using the blinded mode in Rayyan 20 for systematic reviews.Disagreements were resolved through discussion or consulting a third senior researcher (RG).All available data were extracted (GK, CSR), including study characteristics, design, interventions, populations, study methods, and outcomes of significance to the review question and specific objectives (Table 1).Any discrepancies were resolved through discussion or consultation with a senior researcher (RG).
The risk of bias was assessed by five authors (AA, AF, CSR, GK, RB) using Version 2 of the Cochrane risk-of-bias tool for randomised trials (RoB 2) for randomised trials, 21 Risk Of Bias In Non-randomised Studies -of Interventions (ROBINS-I) for observational studies, 22 and the checklist from the National Institutes of Health (NIH) quality assessment tool for beforeafter studies. 23Disagreements were resolved by consensus or discussion with a senior researcher (RG).If information from included studies was not available to answer our primary and secondary outcomes or assess the overall quality of the studies, the corresponding authors were approached.
Three authors (AF, GK, MH) independently assessed the certainty of evidence with Grading of Recommendations Assessment, Development, and Evaluation (GRADE). 24Any disagreements that arose were resolved through consensus.Six clinically relevant outcomes were defined and included in the GRADE table of certainty of evidence (Table 2).

Statistical methods
Only RCTs were considered for meta-analysis.Effect sizes (risk ratios [RRs] for binary outcomes and mean differences for continuous outcomes) were calculated when at least two RCTs reported data for an outcome.
As considerable between-study heterogeneity was detected, a random-effects model was used for analysis of effect sizes.We chose to apply the inverse variance method for continuous outcomes, and ManteleHaenszel for binary outcomes.When median and inter-quartile ranges were reported as summary statistics for continuous outcomes, a quantile estimation method was used to estimate the mean and standard deviation (SD).The between-study heterogeneity was assessed with Higgins and Thompson's I 2 statistic.Statistical tests for funnel plot asymmetry were not performed given the small number of studies per outcome.All statistical computations were performed with R Version 4.0.5. 25

Results
We identified 27,084 articles after deduplication published between 1988 and 2023 (Fig. 1); 15 studies (including 9802 patients) met inclusion criteria.Among these, 10 were RCTs, 26e35 four pre-post studies, 36e39 and one a prospective observational study. 40Ten RCTs were considered eligible for effect size calculation; eight RCTs had an intervention and a control group, one study described two intervention arms and a control group (Steiner and colleagues 34 ), and one study reported the outcomes for two subgroups only (Foran and colleagues 30 ).The two intervention arms in Steiner and colleagues 34 were analysed as separate and independent group comparisons.Similarly, the two subgroups in Foran and colleagues 30 were analysed as two independent intervention arms compared with the control group.The limitation of this approach is outlined in the Limitations section.The meta-analysis of RCTs involved 1070 children, 803 from the operating theatre and 267 from the neonatal intensive care unit.Study characteristics are outlined in Table 1.

Risk of bias assessment
Figure 2 summarises the risk of bias assessment (RoB2, ROBINS-I, pre-post-studies).The overall risk of bias from RCTs was considered to be low in four trials 26,31,33,35 and to have some concerns in six trials.27e30,32,34 Across the trials, the risk of deviations from intended interventions and missing outcome data was deemed low.Some concerns were noted for two trials 27,28 mainly because of inadequate reporting of the randomisation procedure (allocation concealment), concerns over the validity of outcome measurements were noted in three trials 30,32,34 as the assessors were aware of the interventions received by study participants, while in three trials 27,29,30 we were unable to assess adherence to a prespecified study protocol because of unavailability of the latter (outcome-reporting bias).Despite our efforts to contact the authors, we failed to obtain additional clarifying information.The quality rating for the before-after studies assessment was good in one study 36 and fair in three studies.37e39

Meta-analyses
The forest plots of the meta-analyses appear in Figures 3 and 4.
First-pass success rate of tracheal intubation Three RCTs 28,30,33 with a total of 374 patients reported the success rate of first attempt at tracheal intubation, which was achieved in 66.8% (N¼102/155) of the patients with apnoeic oxygenation and 51.5% (N¼87/169) in the control group.The overall pooled analysis (Fig. 3a) showed a higher likelihood of first-pass successful tracheal intubation in the apnoeic oxygenation group compared with the control group (RR 1.27, 95% CI 1.03e1.57,P¼0.04;I 2 ¼0).However, a single study 33 largely dominates the pooled estimate, contributing 84.8% to the pooled effect size.The certainty of evidence was graded low given the small sample size, and as the optimal information size was not met.

Lowest oxygen saturation
Five RCTs 29e33 with six comparisons reported the lowest haemoglobin oxygen saturation during tracheal intubation for 500 children, 248 children in the apnoeic oxygenation group and 252 in the control group (Fig. 3b).Patients in the apnoeic oxygenation group had a higher oxygen saturation than those in the control group (mean difference 3.6%, 95% CI 0.8e6.5%,P¼0.02).Substantial between-study heterogeneity (I 2 ¼63%) was detected.The certainty of evidence was graded high.

Incidence of hypoxaemia
Three RCTs 26,29,34 with 769 children reported the incidence of hypoxaemia during tracheal intubation (379 children in the apnoeic oxygenation group and 390 children in the control group; Fig. 4a).Overall pooled estimate showed a significant reduction in hypoxaemia incidence (RR 0.24, 95% CI 0.17e0.33,P<0.01).Between-study heterogeneity was substantial (I 2 ¼51%).The certainty of evidence was graded very low because of a wide variance of point estimates across studies, a minimal overlap of confidence intervals, a high I 2 statistic, and as optimal information sizes were not met.More importantly, 99.4% of the pooled effect size is derived from the two group comparisons of a single study. 34

Time to successful intubation
Six RCTs 26,27,29,31,33,34 with seven group comparisons reported the time to successful intubation with a total of 1117 children

NIH quality assessment tool for before-after studies:
politano and colleagues (2019) 36 Napolitano and colleagues (2023) 37 Vukovic and colleagues (2019) 39 Yes Yes Yes No Yes Yes No NR Yes Yes No Fair NR=not reported Criteria D1: Was the study question or objective clearly stated?D2: Were eligibility/selection criteria for the study population prespecified and clearly described?D3: Were the participants in the study representative of those who would be eligible for the test/service/intervention in the general or clinical population of interest?D4: Were all eligible participants who met the prespecified entry criteria enrolled?D5: Was the sample size sufficiently large to provide confidence in the findings?D6: Was the test/service/intervention clearly described and delivered consistently across the study population?D7: Were the outcome measures prespecified, clearly defined, valid, reliable, and assessed consistently across all study participants?Ledbetter and colleagues (1988) 27 Windpassinger and colleagues (2016) 26 Steiner and colleagues (2016) 34 Humphreys and colleagues (2017) 35 Dias and colleagues (2017) 29 Olayan and colleagues (2018) 32 Bruckner and colleagues (2021) 28 Gandhi and colleagues (2021) 31 Hodgson and colleagues (2022) 33 Foran and colleagues (2023)  (553 children with apnoeic oxygenation and 564 in the control group; Fig. 4d).The pooled estimate showed no difference in the intubation success time (mean difference 10.5 s, 95% CI À4.4 to 25.4 s, P¼0.14).Between-study heterogeneity was very high (I 2 ¼96%), and the certainty of evidence was graded moderate as the heterogeneity was largely dominated by a single study. 33

Apnoea time
Two RCTs 26,33 reported apnoea times during paediatric tracheal intubation with a total of 296 patients (148 children with apnoeic oxygenation and 148 in the control group; Fig. 3c).No difference was detected in apnoea times (mean difference 19 s, 95% CI À143 to 182 s, P¼0.38).Between-study heterogeneity was substantial (I 2 ¼60%), and the random effects models demonstrated large uncertainties in the pooled estimates as the two studies differed in their mean estimates and SD by a factor of almost 4.The evidence was not graded as the two studies used different apnoea time definitions.The first study 33 reported on duration of apnoea for intubation, whereas in the other study 26 children were first intubated and then left apnoeic until the onset of desaturation with or without oxygenation.

Adverse events
Two RCTs 28,33 reported adverse events during the tracheal intubation.Hodgson and colleagues 33 reported serious adverse events such as chest compressions or epinephrine administration within 1 h after tracheal intubation in 0 of the 124 patients with apnoeic oxygenation vs 1.6% (N¼2/127) in the control group.Pneumothorax was reported in 1.6% (N¼2/124) patients in the apnoeic oxygenation group and 4.7% (N¼6/127) in the control group.Death within 72 h after tracheal intubation in 0.8% (N¼1/ 124) of patients with apnoeic oxygenation and 2.4% (N¼3/127) in the control group.Bruckner and colleagues 28 reported aborted tracheal intubation because of desaturation, bradycardia, or both in 33% (N¼3/10) of intubation attempts with apnoeic oxygenation and in 69% (N¼20/29) when intubation was attempted without oxygen.Because of variability in methods and reporting across these studies, no meta-analysis was conducted.Oxygen for paediatric intubation -11

Discussion
This systematic review and meta-analysis investigated the effectiveness of apnoeic oxygenation during tracheal intubation in children under 16 yr of age.Despite the low certainty of evidence according to GRADE, apnoeic oxygenation was associated with a higher probability of first-pass tracheal intubation success and a reduced number of intubation attempts for each patient.Regardless of the method of administration, apnoeic oxygenation reduced the incidence of hypoxaemia when compared with no oxygen administration.Apnoeic oxygen administered during tracheal intubation  improved and stabilised respiratory and haemodynamic variables and facilitated overall airway management.First-attempt success is critical in paediatric airway management, as adverse events are directly associated with the overall number of intubation attempts. 2,41,42Our findings indicate that administration of oxygen reduces the overall number of attempts by improving first attempt success rate.This is highly important for neonates, small infants, and children with limited cardiopulmonary reserve, as they are more prone to rapid oxygen desaturation.These patients often have higher oxygen consumption, lower closing capacity, low functional residual capacity, and increased risk of airway collapse compared with older children. 43he main obstacle of tracheal intubation in children is often the short apnoea time before severe arterial oxygen desaturation, which is more severe in neonates, infants, and those with severe comorbidities. 12Thus, methods to extend the safe apnoea time are highly desirable.Across the studies included in this systematic review, oxygen delivery techniques, flow rates, and significant hypoxaemia levels were heterogeneous.Nevertheless, in eight RCTs 26e29,31,33e35 apnoeic oxygenation during intubation prolonged the safe apnoea time, decreased the incidence of hypoxaemia, and reduced the number of intubation attempts or increased first-pass success without physiological instability.These findings are supported by five observational or pre-post studies 36e40 that were not included in our meta-analyses but showed a benefit of apnoeic oxygenation during tracheal intubation with fewer adverse events, better intubation conditions, and reduction of hypoxaemia.Maintaining adequate oxygenation during tracheal intubation is crucial in neonates and infants to prevent hypoxaemia and the associated complications. 10,12,44,45n the one hand, using oxygen at high concentrations in premature babies and infants can lead to complications triggered by oxidative stress, including bronchopulmonary dysplasia 46 or severe retinopathy of prematurity. 47Administering supplemental oxygen to children with a cyanotic congenital heart disease might lead to perfusion mismatch, resulting in haemodynamic instability. 48However, hypoxaemia is also harmful, and haemoglobin oxygen saturation values of 85e89% in premature babies increase the risk of death before hospital discharge. 49In addition to potential harm with oxygen, there is also the risk of barotrauma.Hodgson and colleagues 33 diagnosed pneumothorax more often in patients without apnoeic oxygenation and Napolitano and colleagues 37 showed fewer adverse events, including pneumothorax, in an observational study.Causality remains unclear, but rescue facemask ventilation during intubation attempts might pose a risk of barotrauma rather than apnoeic oxygenation.High-flow nasal oxygen administration was reported to reduce the risk of pneumothorax compared with continuous positive airway pressure for respiratory support in preterm infants. 50In summary, the relatively short period of potential hyperoxia during intubation with apnoeic oxygenation appears to outweigh the benefits of avoiding hypoxaemia

Limitations
The significant heterogeneity in study designs of the included trials constitutes a major limitation together with the limited number of included patients and studies.A further limitation is our assumption of the independence of different treatment arms 34 or subgroups. 30It is likely that the true uncertainty of the pooled effect sizes (i.e. the confidence intervals), might be wider than reported in this study.
Avoiding hypoxaemia and keeping a normal oxygen saturation is the first step of a cascade of events that might lead to serious adverse events such as bradycardia or cardiac arrest.The aggregated evidence from this systematic review confirms that apnoeic oxygenation during airway management reduces the incidence of hypoxaemia, but the strength of evidence remains low.This is partly as a result of the exclusion of several studies investigating apnoeic oxygenation because of the lack of a control group with no intervention. 15,51,52Moreover, only two RCTs 28,33 reported data on adverse events, both of which reported a higher incidence of adverse events in children with no apnoeic oxygen administration.However, data were insufficient and not consistently reported to perform a meta-analysis.This is also supported by the findings of the observational studies by Napolitano and colleagues. 36,37hey found an association between oxygen administration and a lower incidence of severe adverse events during tracheal intubation but not a reduction in tracheal intubation attempts or severe peri-intubation hypoxaemia.

Unanswered questions and future research
This systematic review illustrates a high variability in the choice of outcome measures across the included studies.Future study designs on paediatric airway management should standardise the choice of outcomes with actual clinical relevance, allowing better comparison and meta-analysis of studies.First-pass success rate of tracheal intubation and adverse events should be included in any future research as they represent the clinical outcomes of highest relevance.The optimal technique for oxygen delivery during intubation, oxygen flow rate, and concentration of oxygen to be administered during oxygenation have yet to be determined.As financial resources in healthcare and environmental consciousness are increasingly important, cost-effectiveness and environmental impact analysis should be included in future studies.
Severe hypoxaemia, severe bradycardia, cannot intubate and cannot ventilate, emergency front of neck access, cardiac arrest, and death were rarely reported in the included studies, given the relatively small sample sizes of the included individual studies.

Conclusions
This systematic review provides evidence for improved firstattempt intubation success rate during tracheal paediatric intubation with apnoeic oxygenation.Apnoeic oxygenation during tracheal intubation in children decreases the risk of oxygen desaturation.However, as the included data were heterogeneous, more high-quality randomised controlled trials are warranted with a well-defined core set of outcome variables.

Fig 1 .
Fig 1. Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) 2020 flow diagram of study selection.ICTRP, International Clinical Trials Registry Platform.

Criteria D1 :
Bias arising from the randomisation process.D2: Bias because of deviations from the intended interventions.D3: Bias because of missing outcome data.D4: Bias in measurement of the outcome.D5: Bias in the selection of the reported result.

Table 1
Summary of the findings for the included studies and for the observational and pre-post trials.aOR, adjusted odds ratio; ETT, ; int., intervention group; NHF, nasal high flow; PICU, paediatric intensive care unit; RR, risk ratio; Risk of bias methodological quality tools used:* RoB 2: A revised Cochrane risk-of-bias tool for randomised trials.y ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions.z checklist from the NIH Quality Assessment Tool for Before-After (Pre-Post) Studies.

Table 2
Grading of Recommendations Assessment, Development, and Evaluation (GRADE) table of certainty of evidence for the main outcomes.The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).CI, confidence interval; RR, risk ratio.*Optimal information size is not met -when the dominant study (in terms of sample size) is removed as sensitivity analysis.y Wide variance of point estimates across studies, minimal overlap of confidence intervals, high I-squared statistic (>50%) z Optimal information size not met -in particular when the dominant study (in terms of sample size) is removed as sensitivity analysis.¶ Wide variance of point estimates across studies, minimal overlap of confidence intervals, high I-squared statistic (60%) x Very high I-squared statistic (96%), however, this heterogeneity is largely dominated by a single study.

Identification Included Screening Identification of studies via databases and registers Identification of studies via other methods
reported the incidence of bradycardia with 306 children (146 in the apnoeic oxygenation group and 160 in the control group; Fig.4b).No reduction was detected in the