Heliox for prevention of morbidity and mortality in ventilated newborn infants

Objectives

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

To determine the effect of heliox on the rates of mortality, neurodevelopmental outcome, and chronic lung disease in invasively ventilated infants.

Background

Helium, used in combination with oxygen (heliox), facilitates laminar flow in the airways, and offers less resistance to flow due to its low density (one‐seventh the density of nitrogen (Morgan 2014)). Heliox decreases the pressure and work of breathing to ventilate the lung, and may be a useful adjunct in the management of people with airway obstruction, respiratory syncytial virus bronchiolitis, and respiratory distress (Martinon‐Torres 2012).

Description of the condition

Respiratory distress syndrome is the commonest form of respiratory disease in premature neonates, and correlates with structural and functional pulmonary immaturity (Bancalari 2012). The term refers to preterm neonates with oxygen or respiratory support requirements after delivery (Gomella 2020). Bronchopulmonary dysplasia, or chronic lung disease, describes the ongoing requirement for oxygen or respiratory support at 36 weeks' corrected gestational age (Jobe 2001). Mainstays of treatment include oxygen supplementation and ventilation, both invasive and non‐invasive (Sweet 2019). Prolonged mechanical ventilation is associated with increased incidence of chronic lung disease, neurodevelopmental delay, and death (Jobe 1998; Nowadsky 2009). The advent of antenatal steroids, surfactant, and continuous positive airway pressure have improved respiratory outcomes, but further decreasing the duration of ventilation and inflammatory changes may improve neonatal outcomes (Dani 2016; Lopez 2013). In addition, minimising oxygen toxicity and the work of breathing using helium‐oxygen rather than nitrogen‐oxygen may allow the use of lower oxygen concentrations, reducing oxidative damage (Schulzke 2010).

Description of the intervention

Routinely, supplemental oxygen is delivered as an air‐oxygen mixture. However, there are theoretical benefits of using a helium‐oxygen mixture as an alternative to air‐oxygen mixtures for respiratory support. Helium was first discovered by Janssen and Lockyer in 1868 (Kean 2021). It is a colourless, odourless, noble gas that does not react with biological tissues or compounds. Since helium is non‐toxic and biologically inert, it can safely be mixed with oxygen. The helium‐oxygen mixture may be superior to an air‐oxygen mixture. This systematic review will concentrate on the combination of helium and oxygen, known as heliox.

The composition of air is 21% oxygen, 79% nitrogen, and trace amounts of other gases. Heliox is a mixture of 21% oxygen and 79% helium, which is lighter than air or oxygen. Carbon dioxide diffuses more rapidly through a helium‐oxygen mixture (heliox) than through a nitrogen‐oxygen mixture. When supplemental oxygen is inspired, there is a reciprocal effect on inspired nitrogen concentration. The higher the concentration of helium, the lower the fraction of inspired oxygen (FiO₂), and the less dense the gas mixture. Therefore, the major impediment to heliox’s use is hypoxia, because of the limited FiO₂ of 40% that can be achieved, due to a high concentration of helium versus oxygen. Generally, heliox is available in mixtures of 80% oxygen and 20% helium (80:20 heliox) or 70% oxygen and 30% helium (70:30 heliox); however, heliox can be blended with other gases (Hess 2006).

Heliox has a similar viscosity to air but a significantly lower density. Flow of gas through the airway comprises laminar flow, transitional flow, and turbulent flow. The tendency for each type of flow is described by the Reynolds number, a dimensionless number used to determine the likelihood that a gas will develop a laminar or turbulent flow pattern. Heliox's low density produces a lower Reynolds number, and hence, a higher probability of laminar flow for any given airway. Laminar flow tends to generate less resistance than turbulent flow. Heliox generates less airway resistance than air, and thereby, requires less mechanical energy to ventilate the lungs. It does this by two mechanisms: (a) increased tendency to laminar flow; heliox facilitates laminar flow in the airways and offers less resistance to flow due to its low density (one‐seventh that of nitrogen), and (b) reduced resistance in turbulent flow; heliox is less likely to generate turbulent, inefficient flow in situations where airway size is reduced, such as the constricted airways seen in asthma. Indeed, helium‐oxygen gas mixtures have been used in the management of human respiratory disorders, such as asthma and chronic obstructive pulmonary disease, since 1935 (Barach 1935). Methods of administration have progressed, and heliox can now be delivered by several routes, including ventilator, incubator, nasal prongs, and head box (Myers 2003; Singhaus 2006).

Animal models have shown evidence to support the use of, and effects of heliox. Long‐term heliox incubator exposure has been shown to be safe in a randomised controlled trial of neonatal rabbits. There were no differences in growth parameters between those rabbits nursed in heliox versus rabbits nursed in air (Singhaus 2007). In neonatal pigs, heliox was found to reduce tidal volumes with a concomitant therapeutic benefit of attenuating lung inflammation, by reducing mechanical and oxidative stress in the clinical management of acute lung injury (Nawab 2005). A study on ventilated rats showed that neutrophil infiltration, interstitial oedema and intra‐alveolar oedema were all reduced in the group treated with heliox when compared to the control group (Yilmaz 2013).

Beneficial effects have been observed in adults and children with asthma, chronic obstructive pulmonary disease, bronchiolitis, bronchopulmonary dysplasia, and upper airway obstruction (Chevrolet 2001). In adults, the use of helium in the immediate post‐extubation period decreases inspiratory effort and improves comfort (Jaber 2001). In children, heliox has been described as a potentially useful adjunct in the management of children with airway obstruction (Myers 2003). In a consecutive series of children with acute upper airway obstruction, heliox reduced the need for intubation (Connolly 2001). However, although it has been shown to improve clinical scores in infants and children with upper and lower airway respiratory insufficiency, it did not obviate the need for non‐invasive or invasive mechanical ventilation (Iglesias‐Fern 2007).

The properties of heliox provide advantages in situations where there is reduced airway size, as seen in the small airways of premature infants (Jolliet 2003). A pilot time‐series study showed heliox to be well tolerated in mechanically ventilated newborns with broncho pulmonary dysplasia (Szczapa 2014), and in neonates with meconium aspiration syndrome (Szczapa 2011). A multicentre, randomised, blinded controlled trial of 319 infants with bronchiolitis, showed that heliox was effective if delivered via a tight‐fitting, non‐rebreathing face mask or continuous positive airway pressure, but not via a nasal cannula at conventional flow rates (Chowdhury 2013).

The practicalities of administering heliox therapy remained a significant barrier to both its use in NICU, and to studying its effectiveness in high‐powered randomised controlled trials (RCTs). Previously, one of the main issues was equipment, and methods to safely deliver the helium‐oxygen mix. With modern ventilators, this barrier has reduced, but the cost impact remains significant. In the USA, heliox is estimated to cost approximately 20% to 40% more than medical oxygen, which would usually account for less than 1% of the total NICU costs. When the authors looked internationally, this estimated increase in costs was significantly higher, for example, an increase in spending of more than 20‐fold compared to basic oxygen. (Szczapa 2022).

How the intervention might work

Heliox has been used in infant studies for the treatment of respiratory syncytial virus bronchiolitis (Cambonie 2006; Chowdhury 2013; Gross 2000; Martinon‐Torres 2003a; Panitch 2003), and respiratory distress (Elleau 1993), with variable success. In adult and paediatric studies, heliox reduced required ventilating pressures (Jolliet 2003), and inspired oxygen concentrations, reduced arterial carbon dioxide levels (Oca 2002), inspiratory effort (Barach 1935), and the work of breathing (Adb‐Allah 2003), improved comfort scores (Williams 2004), and reduced failure rates after extubation, compared with normal nitrogen‐oxygen mixes (Jaber 2001). However, Cochrane Reviews of heliox used in non‐intubated people with acute asthma (Rodrigo 2006), and exacerbations of chronic obstructive pulmonary disease (Rodrigo 2001), did not have enough evidence to either support or refute routine use of heliox.

Heliox has been found to be safe and well tolerated by both a systematic review (Rodrigo 2003), and a cross‐over study of volunteers (Oei 2012). There are a number of reported adverse effects following the use of heliox. Hypoxia has been reported with heliox in preterm infants who have a history of bronchopulmonary dysplasia and subglottic stenosis (Butt 1985). It was hypothesised that hypoxia in preterm infants secondary to heliox administration was related to the reduction of lung volume and the increase of intrapulmonary shunt (Martinon‐Torres 2003b), and if peak end expiratory pressure was less than 1 cm to 2 cm H₂O, alveolar collapse may occur (Ravenel 1986). Unless the FiO₂ requirement is less than 40%, the person is unlikely to benefit from the small quantity of helium that could be mixed with the inhaled gas (Tassaux 1999). In addition, tidal volume measurements can be altered by the use of heliox in a ventilator calibrated for an air‐oxygen mix, and usually requires a change to pressure‐controlled ventilation and the use of correction factors to accurately measure expired volumes (Tassaux 1999). Hypothermia has been associated with hood administration of heliox to infants. Heliox must be used with caution because of its high thermal conductivity and the consequent risk of hypothermia when the gas temperature is less than 36 °C, especially when heliox is administered for long periods. The risk of hypothermia can be avoided with adequate warming and humidification of the heliox, using standard devices (Martinon‐Torres 2002).

Why it is important to do this review

There is limited experience with the therapeutic use of helium‐oxygen mixtures in the preterm infant. Helium‐oxygen mixtures have been used predominantly in pulmonary function testing (Hentschel 2001; Poets 1996). Compared with pure oxygen, heliox may be a suitable and safer alternative for functional residual capacity measurements, with the nitrogen washout technique in preterm infants who are breathing low concentrations of inspired oxygen, but still at risk of retinopathy of prematurity (Poets 1996). During heliox breathing, preterm infants with bronchopulmonary dysplasia have a significant decrease in pulmonary resistance, resistive work of breathing, and mechanical power of breathing, whereas ventilation remains unchanged. Breathing a lower density gas mixture (heliox) may have therapeutic value, by decreasing the demands on the respiratory muscles and the caloric requirements for breathing. Therefore, this modality may reduce potential respiratory muscle fatigue, and make additional calories available for growth and recovery in the preterm infant with bronchopulmonary dysplasia (Wolfson 1984). The potential for heliox as a therapy to reduce the reintubation rate in this at‐risk group, minimising airway trauma and subsequent problems, has never been investigated.

Objectives

To determine the effect of heliox on the rates of mortality, neurodevelopmental outcome, and chronic lung disease in invasively ventilated infants.

Methods

Criteria for considering studies for this review

Types of studies

We will include randomised controlled trials (RCTs), quasi‐randomised controlled trials, and cluster‐randomised trials.

Types of participants

We will include preterm (less than 37 weeks' gestation) and term infants (37 weeks' gestation or older) requiring invasive ventilation, with a fraction of inspired oxygen (FiO₂) more than 21%.

Types of interventions

We will include any dose of ventilator‐delivered helium‐oxygen (heliox) mix. We will include all types of invasive ventilation, with no limits of duration or frequency of heliox administration.

The comparison group will be controls who were invasively ventilated for any length of time, but did not receive any dose of helium‐oxygen mix.

Types of outcome measures

We will look at primary outcomes, as listed below, for evidence of a direct impact of the outlined intervention on respiratory function.

We will also include secondary outcomes, for association between the intervention and improvement or prevention of organ dysfunction, or general well‐being of the neonate.

Primary outcomes

1. Death at discharge

2. Chronic lung disease/bronchopulmonary dysplasia (oxygen dependency at 36 weeks' postmenstrual age) (Gomella 2020)

3. Adverse events, including hypoxia and hypothermia

4. Long‐term neurodisability (cerebral palsy, developmental delay with Bayley or Griffith assessment more than two standard deviations below the mean, intellectual impairment (intelligence quotient (IQ) more than two standard deviations below the mean), blindness (vision less than 6/60 in both eyes), or sensorineural deafness requiring amplification)

Secondary outcomes

1. Duration of mechanical ventilation via the endotracheal tube, in days

2. Duration of respiratory support via endotracheal tube or continuous positive airway pressure, in days

3. Intraventricular haemorrhage (any (Calisici 2015))

4. Severe intraventricular haemorrhage (grade III or IV)

5. Success of extubation, defined as not requiring re‐intubation within 72 hours. The indications for re‐intubation were defined a priori: (a) more than six episodes of apnoea requiring stimulation in six hours, or more than one significant episode of apnoea requiring bag and mask ventilation; (b) respiratory acidosis (partial pressure of carbon dioxide in arterial blood (PaCO₂) more than 65 mmHg, or 8.5 kPa and pH less than 7.25); (c) FiO₂ more than 60% to maintain SpO₂ in the target range (90% to 95% (Kamlin 2006))

6. Duration of supplemental oxygen therapy, in days

7. Pulmonary function testing

8. Pneumothorax

9. Pulmonary interstitial emphysema

10. Culture positive septicaemia

11. Hypothermia (mild hypothermia 36 ^oC to 36.4 ^oC, moderate 32 ^oC to 35.9 ^oC, severe less than 32 ^oC) according to World Health Organization (WHO) guidelines (McCall 2018)

12. Necrotising enterocolitis (stage 2 Bell or higher (Holman 1997))

13. Duration of neonatal intensive care unit (NICU) stay, in days

14. Duration of hospital stay, in days

15. Patent ductus arteriosus (PDA) requiring treatment (medical or surgical)

16. Cost per cylinder used of helium

Search methods for identification of studies

Electronic searches

In consultation with the authors, the Cochrane Neonatal Information Specialist developed a draft search strategy for MEDLINE Ovid (Appendix 1). This strategy will be peer‐reviewed by an Information Specialist, using the PRESS Checklist (McGowan 2016). We will search the following databases without language, publication year, publication type, or publication status restrictions. We will use methodological filters to identify trials and systematic reviews.

1. Cochrane Central Register of Controlled Trials (CENTRAL)

2. MEDLINE Ovid and Epub Ahead of Print, In‐Process, In‐Data‐Review & Other Non‐Indexed Citations and Daily (1946 to current)

3. Embase Ovid (1974 to current)

We will document the searches in sufficient detail to inform a study flow (PRISMA) diagram (Liberati 2009; Moher 2009).

Searching other resources

We will identify trial registration records using CENTRAL, and by independent searches of the US National Library of Medicine (clinicaltrials.gov); and the World Health Organization International Clinical Trials Registry Platform (ICTRP: www.who.int/clinical-trials-registry-platform/the-ictrp-search-portal).

We will identify conference abstracts using CENTRAL, Embase, and the European Society for Pediatric Research, from 1997 to the most recent year available.

We will screen the reference lists of included studies and related systematic reviews for studies not identified by the database, trial registry, and conference abstract searches.

Data collection and analysis

We will collect data, including the method of randomisation, blinding, intervention, stratification, and whether the trial was single or multicentre for each included study. We will analyse the trial participants, based on gestation, birth weight, ventilation support requirement, clinical diagnoses, and the clinical outcomes outlined in the study. Specifically, we will focus on the outcomes included in our primary and secondary outcomes lists.

Selection of studies

We will include all randomised and quasi‐randomised controlled trials that fulfil the selection criteria described in the previous section. We will download all titles and abstracts, and remove duplicates, to a reference management software program. Two review authors (LR, EM) will independently review the results of the search and select studies for inclusion. We will resolve any disagreement by discussion or by involving a third review author.

We will document the reasons for excluding studies during review of full texts in a 'Characteristics of excluded studies' table. We will collate multiple reports of the same study so that each study, rather than each report or reference, is the unit of interest in the review; related reports will be grouped under a single study. We will also provide any information we can obtain about ongoing studies. We will record the selection process in sufficient detail to complete a PRISMA flow diagram (Liberati 2009; Moher 2009).

Data extraction and management

We will use a data extraction form specifically designed for this review. Two review authors (LR, EM) will independently extract, assess, and code all data for each included study.

We will extract the following characteristics from each included study.

• Administrative details: study author(s); published or unpublished; year of publication; year in which study was conducted; presence of vested interest by study authors; details of other relevant papers cited

• Study characteristics: study registration, study design type, study setting, number of study centres and location; informed consent; ethics approval, details of any 'run‐in' period (if applicable), completeness of follow‐up (e.g. greater than 80%)

• Participants: number randomised, number lost to follow‐up/withdrawn, number analysed, mean gestational age (GA), GA age range, mean chronological age (CA), CA age range, sex, severity of condition, diagnostic criteria, inclusion criteria and exclusion criteria

• Interventions: initiation, dose, and duration of administration

• We will collect information on the clinical outcomes as specified in Types of outcome measures, including the impact of the intervention on respiratory function as a primary outcome, and any other statistically significant associated change in neonatal well‐being or organ dysfunction.

We will resolve any disagreements by discussion.

We will describe ongoing studies identified by our search and document available information such as the primary author, research question(s), methods, and outcome measures, together with an estimate of the anticipated reporting date in the 'Characteristics of ongoing studies' table.

Should any queries arise, or in cases for which additional data are required, we will contact study investigators/authors for clarification.

We intend to perform an intention‐to‐treat analysis, including all participants of the included studies.

We will replace any standard error of the mean (SEM) by the corresponding standard deviation (SD).

For each study, one review author (LR) will enter final data into RevMan Web 2023, which will be checked for accuracy by a second review author (EM). We will resolve any discrepancies through discussion with the other members of the authorship team (LR, TH, EM, AB).

Assessment of risk of bias in included studies

We will use RoB 2 to assess the risk of bias in the included studies (Sterne 2019). The outcomes that we will assess for each study are specified in Summary of findings and assessment of the certainty of the evidence.

Two review authors (LR, EM) will independently assess the risk of bias (low, high, or some concerns) in all included studies. In case of discrepancies amongst their judgements and inability to reach consensus, we will consult a third review author (AB) to reach a final decision. We will assess the following types of bias as outlined in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022).

1. Bias arising from the randomisation process

2. Bias due to deviations from intended interventions

3. Bias due to missing outcome data

4. Bias in measurement of the outcome

5. Bias in selection of the reported result

To address these types of bias, we will use the signalling questions recommended in RoB 2 (see Appendix 2), and make a judgement using the following options.

1. Yes: if there is firm evidence that the question was fulfilled in the study (i.e. the study was at low or high risk of bias given the direction of the question)

2. Probably yes: a judgement was made that the question was fulfilled in the study (i.e. the study was at low or high risk of bias given the direction of the question)

3. No: if there was firm evidence that the question was unfulfilled in the study (i.e. the study was at low or high risk of bias given the direction of the question)

4. Probably no: a judgement was made that the question was unfulfilled in the study (i.e. the study was at low or high risk of bias given the direction of the question)

5. No information: if the study report provided insufficient information to allow any judgement

We will use the algorithms proposed by RoB 2 to assign each domain one of the following levels of bias: low risk of bias; some concerns; high risk of bias. This allows the review authors to derive an overall risk of bias rating for each outcome in each study.

1. Low risk of bias: we judged the trial at low risk of bias for all domains for this result.

2. Some concerns: we judged the trial to raise some concerns in at least one domain for this result, but not at high risk of bias for any domain.

3. High risk of bias: we judged the trial at high risk of bias in at least one domain for the result, or we judged the trial to have some concerns for multiple domains in a way that substantially lowered confidence in the results.

We will use the RoB 2 Excel tool to implement RoB 2 (available at www.riskofbias.info/).

Measures of treatment effect

We will perform statistical analyses using Review Manager software (RevMan Web 2023). We will analyse dichotomous data using risk ratio (RR), risk difference (RD), and the number needed to treat for an additional beneficial outcome (NNTB), or the number needed to treat for an additional harmful outcome (NNTH), for those outcomes where the relative risk or risk difference are statistically significant. The 95% confidence interval (CI) will be reported for all estimates.

We plan to analyse continuous data using the mean difference, or the standardised mean difference (SMD) to combine studies that measure the same outcome but use different methods.

Unit of analysis issues

There are unlikely to be issues with unit of analysis, which will be the individual neonate, randomised to receive the intervention.

For cluster‐randomised trials, we will assess the unit of analysis for each study, and describe the study design; the likely allocation of the intervention made at a group level will be all inpatients in a particular unit. We will pay specific attention to potential confounding factors that may impact the results, unrelated to the intervention, including differences in practice across units, or heterogeneity in cohorts of neonates involved, not comparing like‐with‐like.

If a relevant trial has multiple arms that are compared against the same control, for example, varying doses of helium‐oxygen mix, which are included in the same meta‐analysis, we will either combine them to treat them as one entity in a direct comparison, or select one pair of interventions to analyse and exclude the others.

Dealing with missing data

For included studies, we intend to explore the impact of including studies with high levels of missing data in the overall assessment of treatment effect by using sensitivity analysis.

All outcome analyses will be on an intention‐to‐treat basis, that is, we will include all participants randomised to each group in the analyses. The denominator for each outcome in each study will be the number randomised minus any participants whose outcomes were known to be missing.

Assessment of heterogeneity

Part of our analysis will be based on the degree of heterogeneity present between the included studies, both in the clinical group and the methodology of data collection and analysis. We will examine heterogeneity between trials by inspecting the forest plots, and quantifying the impact of heterogeneity using the I² statistic (Higgins 2003). If noted, we plan to explore the possible causes of statistical heterogeneity using prespecified subgroup analysis, as listed below. To assess the significance of heterogeneity, we will analyse the P value and the degree of heterogeneity involved, ranging from slight and likely unimportant heterogeneity (0% to 40%), to moderate (30% to 60%), substantial (50 to 90%) and considerable degree of heterogeneity (more than 75%). We will also take into account that the statistical power of our planned methods of heterogeneity analysis may be reduced if the number of included studies is small (Deeks 2022).

Assessment of reporting biases

We will assess possible publication bias and other biases using symmetry/asymmetry of funnel plots, if we have sufficient data to develop them.

For included studies that were recently performed (and therefore prospectively registered), we will assess reporting bias by comparing the stated primary and secondary outcomes with the reported outcomes. Where study protocols are available, we will compare these to the full publications to determine the likelihood of reporting bias. Studies using the interventions in a potentially eligible infant population but not reporting on any of the primary and secondary outcomes will be documented in the 'Characteristics of included studies' tables. If such discrepancies are found, we plan to contact the primary investigators to obtain missing outcome data on outcomes prespecified at trial registration.

We will use the funnel plots to screen for publication bias where there are a sufficient number of studies (> 10) reporting the same outcome. If publication bias is suggested by a significant asymmetry of the funnel plot on visual assessment, we will incorporate this in our assessment of certainty of evidence (Egger 1997). If our review includes fewer than 10 studies eligible for meta‐analysis, the ability to detect publication bias will be largely diminished, and we will simply note our inability to rule out possible publication bias or small study effects.

Data synthesis

If we identify multiple studies that we consider sufficiently similar, we will perform meta‐analysis and undertake the analysis using Cochrane Review Manager software (RevMan Web 2023). We will use the Mantel‐Haenszel method for estimates of RR and RD. We do not plan to include continuous outcomes in this review, but we will use the inverse variance method, if they are.

We will use the fixed‐effect model for all meta‐analyses.

If we judge meta‐analysis to be inappropriate, we will analyse and interpret individual studies separately. If there is evidence of clinical heterogeneity, we will try to explain this based on the different study characteristics and subgroup analyses.

Subgroup analysis and investigation of heterogeneity

The subgroups we intend to analyse are:

1. preterm infants of less than 37 weeks' gestation;

2. infants with evolving or established chronic lung disease.

Sensitivity analysis

If we identify substantial heterogeneity, we will conduct sensitivity analysis to determine whether the findings are affected by the inclusion of only those studies considered to have used adequate methodology (at low risk of selection, performance, and reporting bias). We will report the results of sensitivity analyses for primary outcomes only. Given there is no formal statistical test that can be used for sensitivity analysis, we will make informal comparisons between the different ways of estimating the effect under different assumptions. We will not consider changes in P value.

We will report sensitivity analysis results in tables rather than forest plots.

Summary of findings and assessment of the certainty of the evidence

Two review authors (LR, EM) will independently use the GRADE approach, as outlined in the GRADE Handbook, to assess the certainty of evidence for the following (clinically relevant) outcomes (Schünemann 2013).

1. Death at discharge

2. Chronic lung disease/bronchopulmonary dysplasia

3. Adverse events, including hypoxia and hypothermia

4. Long‐term neurodisability

5. Duration of mechanical ventilation via the endotracheal tube

6. Duration of respiratory support via endotracheal tube or continuous positive airway pressure

7. Duration of neonatal intensive care unit stay

We will consider evidence from RCTs as high certainty, downgrading the evidence one level for serious (or two levels for very serious) limitations based upon the following: design (risk of bias), consistency across studies, directness of the evidence, precision of estimates, and presence of publication bias. We will use GRADEpro GDT software to create a summary of findings table to report the certainty of the evidence (GRADEpro GDT).

The GRADE approach results in an assessment of the certainty of a body of evidence in one of the following four grades.

1. High: we are very confident that the true effect lies close to that of the estimate of the effect.

2. Moderate: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.

3. Low: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect.

4. Very low: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

Appendices

Appendix 1. MEDLINE search strategy

	Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process, In‐Data‐Review & Other Non‐Indexed Citations, Daily and Versions(R) 1946 to August 25, 2021
#	Searches	Results
1	Helium/	5072
2	(helium$ or heliox*).ti,ab,kw,kf.	12560
3	or/1‐2 [Helium]	14509
4	exp infant, newborn/	633289
5	(baby* or babies or infant or infants or infant? or infantile or infancy or low birth weight or low birthweight or neonat* or newborn* or new born or new borns or newly born or premature or prematures or prematurity or preterm or preterms or pre term or preemie or preemies or premies or premie or VLBW or LBW or ELBW or NICU).ti,ab,kw,kf.	940089
6	or/4‐5 [Filter: Neonatal Population 2021‐‐MEDLINE]	1217194
7	randomized controlled trial.pt.	541756
8	controlled clinical trial.pt.	94352
9	(randomized or randomised).ti,ab.	686459
10	placebo.ab.	220802
11	drug therapy.fs.	2365410
12	randomly.ab.	364458
13	trial.ab.	565277
14	groups.ab.	2237651
15	(quasirandom* or quasi‐random*).ti,ab.	5193
16	exp animals/ not humans/	4878405
17	(or/7‐15) not 16 [RCT Filter‐Sensitivity Max ]	4464368
18	3 and 6 [Helium & Neonate]	292
19	3 and 6 and 17 [Helium & Neonate & RCT]	77
20	systematic review.pt.	166307
21	(systematic adj2 review).ti.	163036
22	meta analysis/	140467
23	(meta‐analysis or metaanalysis).ti,ab,kw.	184218
24	(cochrane or systematic review?).jw.	18672
25	overview of reviews.ti.	85
26	or/20‐25 [SR filter]	322525
27	3 and 6 [Helium & Neonate‐‐Results before filters]	292
28	3 and 6 and 17 [Helium & Neonate & RCT]	77
29	3 and 6 and 26 [Helium & Neonate & SR]	10

Appendix 2. RoB 2 tool

RoB 2 tool (Higgins 2019)

Risk of bias in the randomisation process

1.1 Was the allocation sequence random?

Y/PY/PN/N/NI

1.2 Was the allocation sequence concealed until participants were enroled and assigned to interventions?

Y/PY/PN/N/NI

1.3 Did baseline differences between intervention groups suggest a problem with the randomisation process?

Y/PY/PN/N/NI

Risk of bias from deviations from planned intervention

2.1. Were participants aware of their assigned intervention during the trial?

Y/PY/PN/N/NI

2.2. Were carers and people delivering the interventions aware of participants' assigned intervention during the trial?

Y/PY/PN/N/NI

2.3. If Y/PY/NI to 2.1 or 2.2: Were there deviations from the intended intervention that arose because of the trial context?

NA/Y/PY/PN/N/NI

2.4 If Y/PY to 2.3: Were these deviations likely to have affected the outcome?

NA/Y/PY/PN/N/NI

2.5. If Y/PY/NI to 2.4:
Were these deviations from intended intervention balanced between groups?

NA/Y/PY/PN/N/NI

2.6 Was an appropriate analysis used to estimate the effect of assignment to intervention?

Y/PY/PN/N/NI

2.7 If N/PN/NI to 2.6: Was there potential for a substantial impact (on the result) of the failure to analyse participants in the group to which they were randomised?

NA/Y/PY/PN/N/NI

Risk due to missing data

3.1 Were data for this outcome available for all, or nearly all, participants randomised?

Y/PY/PN/N/NI

3.2 If N/PN/NI to 3.1: Is there evidence that the result was not biased by missing outcome data?

NA/Y/PY/PN/N

3.3 If N/PN to 3.2: Could absence in the outcome depend on its true value?

NA/Y/PY/PN/N/NI

3.4 If Y/PY/NI to 3.3: Is it likely that missingness in the outcome depended on its true value?

NA/Y/PY/PN/N/NI

Risk in the measurement of the outcome

4.1 Was the method of measuring the outcome inappropriate?

Y/PY/PN/N/NI

4.2 Could measurement or ascertainment of the outcome have differed between intervention groups?

Y/PY/PN/N/NI

4.3 If N/PN/NI to 4.1 and 4.2: Were outcome assessors aware of the intervention received by study participants?

NA/Y/PY/PN/N/NI

4.4 If Y/PY/NI to 4.3: Could assessment of the outcome have been influenced by knowledge of intervention received?

NA/Y/PY/PN/N/NI

4.5 If Y/PY/NI to 4.4: Is it likely that assessment of the outcome was influenced by knowledge of intervention received?

NA/Y/PY/PN/N/NI

Risk in reporting results

5.1 Were the data that produced this result analysed in accordance with a prespecified analysis plan that was finalised before unblinded outcome data were available for analysis?

If the researchers’ prespecified intentions are available in sufficient detail, then planned outcome measurements and analyses can be compared with those presented in the published report(s). To avoid the possibility of selection of the reported result, finalisation of the analysis intentions must precede availability of unblinded outcome data to the trial investigators. Changes to analysis plans that were made before unblinded outcome data were available, or that were clearly unrelated to the results (e.g. due to a broken machine making data collection impossible) do not raise concerns about bias in selection of the reported result.

Y/PY/PN/N/NI

Is the numerical result being assessed likely to have been selected, on the basis of the results, from ...

5.2. ... multiple eligible outcome measurements (e.g. scales, definitions, time points) within the outcome domain?

Y/PY/PN/N/NI

5.3 ... multiple eligible analyses of the data?

Y/PY/PN/N/NI

We will use the suggestion algorithm with this ROB 2 tool (Higgins 2019).

KEY

Y: Yes;PY: Probably yes;PN: Probably no;N: No; NI: No information.

Contributions of authors

LR, EM, TH, and AB contributed to the development and design of this protocol.

Sources of support

Internal sources

• Department of Paediatrics, Trinity College Dublin, Ireland

  Home institution of EM, TH and EB

• National Maternity Hospital, Ireland

  Home institution of LR

External sources

• Vermont Oxford Network, USA

  Cochrane Neonatal Reviews are produced with support from Vermont Oxford Network, a worldwide collaboration of health professionals dedicated to providing evidence‐based care of the highest quality for newborn infants and their families.

Declarations of interest

LR has no interests to declare.

EM is a paid Associate Editor‐in‐Chief of Pediatric Research.

TH has no interests to declare.

AB has no interests to declare.

New