Nasal high flow therapy for primary respiratory support in preterm infants

Abstract

Background

Nasal high flow (nHF) therapy provides heated, humidified air and oxygen via two small nasal prongs, at gas flows of more than 1 litre/minute (L/min), typically 2 L/min to 8 L/min. nHF is commonly used for non‐invasive respiratory support in preterm neonates. It may be used in this population for primary respiratory support (avoiding, or prior to the use of mechanical ventilation via an endotracheal tube) for prophylaxis or treatment of respiratory distress syndrome (RDS). This is an update of a review first published in 2011 and updated in 2016.

Objectives

To evaluate the benefits and harms of nHF for primary respiratory support in preterm infants compared to other forms of non‐invasive respiratory support.

Search methods

We used standard, extensive Cochrane search methods. The latest search date March 2022.

Selection criteria

We included randomised or quasi‐randomised trials comparing nHF with other forms of non‐invasive respiratory support for preterm infants born less than 37 weeks' gestation with respiratory distress soon after birth.

Data collection and analysis

We used standard Cochrane Neonatal methods. Our primary outcomes were 1. death (before hospital discharge) or bronchopulmonary dysplasia (BPD), 2. death (before hospital discharge), 3. BPD, 4. treatment failure within 72 hours of trial entry and 5. mechanical ventilation via an endotracheal tube within 72 hours of trial entry. Our secondary outcomes were 6. respiratory support, 7. complications and 8. neurosensory outcomes. We used GRADE to assess the certainty of evidence.

Main results

We included 13 studies (2540 infants) in this updated review. There are nine studies awaiting classification and 13 ongoing studies. The included studies differed in the comparator treatment (continuous positive airway pressure (CPAP) or nasal intermittent positive pressure ventilation (NIPPV)), the devices for delivering nHF and the gas flows used. Some studies allowed the use of 'rescue' CPAP in the event of nHF treatment failure, prior to any mechanical ventilation, and some allowed surfactant administration via the INSURE (INtubation, SURfactant, Extubation) technique without this being deemed treatment failure. The studies included very few extremely preterm infants less than 28 weeks' gestation. Several studies had unclear or high risk of bias in one or more domains.

Nasal high flow compared with continuous positive airway pressure for primary respiratory support in preterm infants

Eleven studies compared nHF with CPAP for primary respiratory support in preterm infants. When compared with CPAP, nHF may result in little to no difference in the combined outcome of death or BPD (risk ratio (RR) 1.09, 95% confidence interval (CI) 0.74 to 1.60; risk difference (RD) 0, 95% CI −0.02 to 0.02; 7 studies, 1830 infants; low‐certainty evidence). Compared with CPAP, nHF may result in little to no difference in the risk of death (RR 0.78, 95% CI 0.44 to 1.39; 9 studies, 2009 infants; low‐certainty evidence), or BPD (RR 1.14, 95% CI 0.74 to 1.76; 8 studies, 1917 infants; low‐certainty evidence). nHF likely results in an increase in treatment failure within 72 hours of trial entry (RR 1.70, 95% CI 1.41 to 2.06; RD 0.09, 95% CI 0.06 to 0.12; number needed to treat for an additional harmful outcome (NNTH) 11, 95% CI 8 to 17; 9 studies, 2042 infants; moderate‐certainty evidence). However, nHF likely does not increase the rate of mechanical ventilation (RR 1.04, 95% CI 0.82 to 1.31; 9 studies, 2042 infants; moderate‐certainty evidence). nHF likely results in a reduction in pneumothorax (RR 0.66, 95% CI 0.40 to 1.08; 10 studies, 2094 infants; moderate‐certainty evidence) and nasal trauma (RR 0.49, 95% CI 0.36 to 0.68; RD −0.06, 95% CI −0.09 to −0.04; 7 studies, 1595 infants; moderate‐certainty evidence).

Nasal high flow compared with nasal intermittent positive pressure ventilation for primary respiratory support in preterm infants

Four studies compared nHF with NIPPV for primary respiratory support in preterm infants. When compared with NIPPV, nHF may result in little to no difference in the combined outcome of death or BPD, but the evidence is very uncertain (RR 0.64, 95% CI 0.30 to 1.37; RD −0.05, 95% CI −0.14 to 0.04; 2 studies, 182 infants; very low‐certainty evidence). nHF may result in little to no difference in the risk of death (RR 0.78, 95% CI 0.36 to 1.69; RD −0.02, 95% CI −0.10 to 0.05; 3 studies, 254 infants; low‐certainty evidence). nHF likely results in little to no difference in the incidence of treatment failure within 72 hours of trial entry compared with NIPPV (RR 1.27, 95% CI 0.90 to 1.79; 4 studies, 343 infants; moderate‐certainty evidence), or mechanical ventilation within 72 hours of trial entry (RR 0.91, 95% CI 0.62 to 1.33; 4 studies, 343 infants; moderate‐certainty evidence). nHF likely results in a reduction in nasal trauma, compared with NIPPV (RR 0.21, 95% CI 0.09 to 0.47; RD −0.17, 95% CI −0.24 to −0.10; 3 studies, 272 infants; moderate‐certainty evidence). nHF likely results in little to no difference in the rate of pneumothorax (RR 0.78, 95% CI 0.40 to 1.53; 4 studies, 344 infants; moderate‐certainty evidence).

Nasal high flow compared with ambient oxygen

We found no studies examining this comparison.

Nasal high flow compared with low flow nasal cannulae

We found no studies examining this comparison.

Authors' conclusions

The use of nHF for primary respiratory support in preterm infants of 28 weeks' gestation or greater may result in little to no difference in death or BPD, compared with CPAP or NIPPV. nHF likely results in an increase in treatment failure within 72 hours of trial entry compared with CPAP; however, it likely does not increase the rate of mechanical ventilation. Compared with CPAP, nHF use likely results in less nasal trauma and likely a reduction in pneumothorax. As few extremely preterm infants less than 28 weeks' gestation were enrolled in the included trials, evidence is lacking for the use of nHF for primary respiratory support in this population.

Plain language summary

Nasal high flow therapy for breathing support in preterm babies

Review question

In preterm babies, what are the benefits and harms of nasal high flow therapy (high flow) when used for breathing support soon after birth, compared with other types of non‐invasive breathing support?

What is respiratory support and how is it treated?

Preterm infants (born before their due date) often require support with their breathing soon after birth. Non‐invasive respiratory support is provided without placing a breathing tube in the baby's windpipe. There are several types of non‐invasive respiratory support. High flow is one type that delivers warm air and oxygen via two small prongs that sit inside the infant's nostrils. Alternatives to high flow include continuous positive airway pressure (CPAP), where continuous pressure (rather than flow) of oxygen is given via larger prongs or a mask, and nasal intermittent positive pressure ventilation (NIPPV) where, in addition to CPAP, inflations of oxygen at a higher pressure are occasionally given.

What did we do?

We searched medical databases for well‐designed studies evaluating the benefits and harms of high flow respiratory support in preterm infants compared to other forms of non‐invasive respiratory support.

What did we find?

We found 13 studies including 2540 preterm babies that compared high flow with other non‐invasive ways of supporting babies' breathing soon after birth. There are nine studies awaiting classification and 13 ongoing studies. The included studies differed in the treatments they compared, the flows of oxygen used, whether CPAP could be used if high flow did not work and the approach to the use of surfactant (a medication used to help prevent the small airways collapsing down) in babies with more severe breathing difficulties.

What did we find?

When used soon after birth in preterm babies, high flow may make little to no difference to death or bronchopulmonary dysplasia (a chronic lung disease in preterm babies) compared with CPAP or NIPPV. High flow probably increases treatment failure compared with CPAP. For example, babies treated with high flow might have needed higher oxygen concentrations or had worse blood test results. CPAP worked better than high flow in around 10 more babies out of every 100. However, high flow probably makes little to no difference to the likelihood of needing intubation (placement of a breathing tube). High flow probably caused less damage to the infant's nose, compared with CPAP or NIPPV and probably reduced the risk of pneumothorax (air in the space between the lung and the chest wall). There were very few extremely preterm infants (born before 28 weeks' gestation) included in these studies. Therefore, we remain unsure about the benefits and harms of high flow soon after birth for extremely preterm infants.

What are the limitations of the evidence?

Overall, we have very low to moderate confidence in these findings. Our confidence is limited because clinicians in the studies knew which treatment babies received; the results varied widely as some studies showed benefit with one type of breathing support while other studies showed benefit with the comparator type of breathing support; and there were low numbers of events for some outcomes making it difficult to compare groups.

How up to date is this evidence?

The search is up to date as of 12 March 2022.

Summary of findings

Summary of findings 1. Nasal high flow compared to continuous positive airway pressure for primary respiratory support in preterm infants.

Nasal high flow (nHF) compared to continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants
Patient or population: preterm infants requiring primary respiratory support Setting: neonatal intensive care units (Australia, China, Iran, Israel, Italy, Korea, Norway, Turkey, USA) Intervention: nHF Comparison: CPAP
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with CPAP	Risk with nHF
Death (before hospital discharge) or BPD (supplemental oxygen/respiratory support at 36 weeks' postmenstrual age if born < 32 weeks' gestation, or 28 days if born ≥ 32 weeks' gestation)	47 per 1000	52 per 1000 (35 to 76)	RR 1.09 (0.74 to 1.60)	1830 (7 RCTs)	⊕⊕⊝⊝ Lowa,b	nHF may result in little to no difference in death or BPD.
Death (before hospital discharge)	23 per 1000	18 per 1000 (10 to 31)	RR 0.78 (0.44 to 1.39)	2009 (9 RCTs)	⊕⊕⊝⊝ Lowa,b	nHF may result in little to no difference in death.
BPD (supplemental oxygen/respiratory support at 36 weeks' postmenstrual age if born < 32 weeks' gestation, or 28 days if born ≥ 32 weeks' gestation)	33 per 1000	37 per 1000 (24 to 58)	RR 1.14 (0.74 to 1.76)	1917 (8 RCTs)	⊕⊕⊝⊝ Lowa,b	nHF may result in little to no difference in BPD.
Treatment failure within 72 hours of trial entry	131 per 1000	223 per 1000 (185 to 270)	RR 1.70 (1.41 to 2.06)	2042 (9 RCTs)	⊕⊕⊕⊝ Moderatec	nHF likely results in an increase in treatment failure within 72 hours of trial entry.
Mechanical ventilation within 72 hours of trial entry	118 per 1000	122 per 1000 (96 to 154)	RR 1.04 (0.82 to 1.31)	2042 (9 RCTs)	⊕⊕⊕⊝ Moderated	nHF likely does not increase mechanical ventilation within 72 hours of trial entry.
Pneumothorax (during assigned treatment)	34 per 1000	22 per 1000 (14 to 37)	RR 0.66 (0.40 to 1.08)	2094 (10 RCTs)	⊕⊕⊕⊝ Moderatea,e	nHF likely results in a reduction in pneumothorax.
Nasal trauma (during assigned treatment)	125 per 1000	61 per 1000 (45 to 85)	RR 0.49 (0.36 to 0.68)	1595 (7 RCTs)	⊕⊕⊕⊝ Moderatef	nHF likely results in a reduction in nasal trauma.
*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). BPD: bronchopulmonary dysplasia; CI: confidence interval; CPAP: continuous positive airway pressure; nHF: nasal high flow; RCT: randomised controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: 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. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aBlinding of nHF and CPAP not possible, but objective outcome assessment therefore not downgraded.
^bDowngraded two levels due to very serious concerns regarding imprecision: wide confidence intervals including clinically important benefit and harm.
^cDowngraded one level due to serious concerns regarding risk of bias. Blinding of nHF and CPAP not possible. Most trials had objective criteria, but some included subjective criteria or allowed the use of surfactant.
^dDowngraded one level due to serious concerns regarding risk of bias. Blinding of nHF and CPAP not possible. Most trials had objective criteria, but some allowed cross‐over to CPAP with nHF failure, or bilevel positive airway pressure/nasal intermittent positive pressure ventilation with CPAP failure.
^eDowngraded one level due to serious concerns regarding imprecision: fewer infants than optimal information size.
^fDowngraded one level due to serious concerns regarding risk of bias: blinding of nHF and CPAP not possible, with risk of bias in outcome assessment.

Summary of findings 2. Nasal high flow compared to nasal intermittent positive pressure ventilation for primary respiratory support in preterm infants.

Nasal high flow (nHF) compared to nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants
Patient or population: preterm infants requiring primary respiratory support Setting: neonatal intensive care units (China, Iran) Intervention: nHF Comparison: NIPPV
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with NIPPV	Risk with nHF
Death (before hospital discharge) or BPD (supplemental oxygen/respiratory support at 36 weeks' postmenstrual age if born < 32 weeks' gestation, or 28 days if born ≥ 32 weeks' gestation)	144 per 1000	92 per 1000 (43 to 198)	RR 0.64 (0.30 to 1.37)	182 (2 RCTs)	⊕⊝⊝⊝ Very lowa,b	nHF may have little to no effect on death or BPD but the evidence is very uncertain.
Death (before hospital discharge)	102 per 1000	80 per 1000 (37 to 173)	RR 0.78 (0.36 to 1.69)	254 (3 RCTs)	⊕⊕⊝⊝ Lowb,c	nHF may result in little to no difference in death.
BPD (supplemental oxygen/respiratory support at 36 weeks' postmenstrual age if born < 32 weeks' gestation, or 28 days if born ≥ 32 weeks' gestation)	118 per 1000	140 per 1000 (78 to 249)	RR 1.19 (0.66 to 2.12)	271 (3 RCTs)	⊕⊕⊝⊝ Lowb,c	nHF may result in little to no difference in BPD.
Treatment failure within 72 hours of trial entry	231 per 1000	210 per 1000 (143 to 308)	RR 1.27 (0.90 to 1.79)	343 (4 RCTs)	⊕⊕⊕⊝ Moderated	nHF may result in little to no difference in treatment failure within 72 hours of trial entry.
Mechanical ventilation within 72 hours of trial entry	199 per 1000	179 per 1000 (109 to 290)	RR 0.91 (0.62 to 1.33)	343 (4 RCTs)	⊕⊕⊕⊝ Moderated	nHF likely results in little to no difference in mechanical ventilation within 72 hours of trial entry.
Pneumothorax (during assigned treatment)	98 per 1000	76 per 1000 (39 to 149)	RR 0.78 (0.40 to 1.53)	344 (4 RCTs)	⊕⊕⊕⊝ Moderatec,e	nHF likely results in little to no difference in pneumothorax.
Nasal trauma (during assigned treatment)	212 per 1000	44 per 1000 (19 to 99)	RR 0.21 (0.09 to 0.47)	272 (3 RCTs)	⊕⊕⊕⊝ Moderatea	nHF likely results in a reduction in nasal trauma.
*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). BPD: bronchopulmonary dysplasia; CI: confidence interval; nHF: nasal high flow; NIPPV: nasal intermittent positive pressure ventilation; RCT: randomised controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: 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. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded one level due to serious concerns regarding risk of bias: blinding of nHF and NIPPV not possible and unclear objectivity of outcome assessment.
^bDowngraded two levels due to very serious concerns regarding imprecision: wide confidence intervals including clinically important benefit and harm.
^cBlinding of nHF and NIPPV not possible, but objective outcome assessment therefore not downgraded.
^dDowngraded one level due to serious concerns regarding risk of bias: blinding of nHF and NIPPV not possible. Most trials had objective criteria, but some included subjective criteria or allowed the use of surfactant.
^eDowngraded one level due to imprecision: fewer infants than optimal information size.

Background

Description of the condition

Many preterm infants require assistance with their breathing after birth. Respiratory support in preterm infants aims to improve gas exchange and minimise apnoea, while limiting short‐ and long‐term respiratory and neurodevelopmental complications. Use of mechanical ventilation via an endotracheal tube and supplemental oxygen increase the risk of bronchopulmonary dysplasia (BPD) (Davidson 2017; Schmölzer 2013). Non‐invasive respiratory support (i.e. without the use of an endotracheal tube) is therefore commonly employed as a first‐line therapy for prophylaxis or treatment of respiratory distress syndrome (RDS) in preterm neonates. Non‐invasive respiratory support may take several forms including 'low flow' nasal cannulae, continuous positive airway pressure (CPAP), nasal intermittent positive pressure ventilation (NIPPV), and nasal high flow (nHF).

Description of the intervention

Until recently, nasal CPAP has been the most common mode of non‐invasive respiratory support in preterm neonates. Nasal CPAP provides continuous distending pressure via binasal prongs or a nasal mask. It is an effective alternative to elective endotracheal intubation in very preterm infants for the treatment of RDS (Morley 2008), and may reduce the risk of BPD (Schmölzer 2013). Alternatives to CPAP include nasal intermittent positive pressure ventilation (NIPPV) and nHF. NIPPV augments CPAP by providing inflations to a set peak pressure; these may be synchronised or non‐synchronised with the infant's own breathing efforts (Lemyre 2017).

nHF delivers heated, humidified air or oxygen, or both, via small tapered binasal prongs, at gas flows of more than 1 litre/minute (L/min). In contrast to the tightly fitting prongs of nasal CPAP, nHF prongs are designed to avoid occlusion of the nares; leak around the prongs serves to avoid excessive pressure generation. Gas flow, rather than pressure, is set by the clinician, and neonatal gas flows are generally 2 L/min to 8 L/min. nHF is an alternative form of non‐invasive respiratory support for preterm infants with RDS. Compared with CPAP, nHF may be preferred by nursing staff (Roberts 2014), and parents (Klingenberg 2014), due to ease of use and comfort of the infant (Osman 2015).

Both CPAP and nHF systems may have adverse effects in newborns. CPAP interfaces may be bulky and skilled nursing care is required to optimally manage infants, and CPAP use is associated with nasal trauma (Imbulana 2018), and air leaks from the lung (pneumothorax) (Ho 2020).

Risk factors for severe nasal trauma in preterm infants include lower gestational age, lower birth weight, and incorrect sizing and positioning of CPAP prongs (Imbulana 2018). nHF may be associated with a lower risk of nasal trauma, compared with CPAP (Imbulana 2018).

Pneumothorax is another recognised complication of CPAP and nHF; the reported incidence of this complication in preterm infants with RDS is variable (Dunn 2011; Morley 2008). Pneumothoraces occur predominantly in surfactant‐deplete preterm infants receiving treatment for RDS, rather than after surfactant treatment or a period of mechanical ventilation, and are likely secondary to overdistension of some areas of the lung. There were early concerns that the use of nHF may lead to an increase in lung overdistension and pneumothorax from unmeasured positive end expiratory pressure (PEEP) (Hegde 2013; Jasin 2008).

The cost‐effectiveness of nHF compared with CPAP is unclear, and is influenced by the cost and lifespan of the devices, the cost of consumables and the efficacy of the treatment, given the associated costs of mechanical ventilation or surfactant therapy in the event of treatment failure (Fleeman 2016; Huang 2018).

How the intervention might work

Several physiological mechanisms may contribute to the efficacy of nHF: provision of continuous distending pressure (Collins 2013a; Wilkinson 2008), improved alveolar ventilation due to nasopharyngeal dead space washout (Dysart 2009; Liew 2020; Sivieri 2017), and reduction of airway resistance and therefore 'work of breathing' (Lavizzari 2014; Saslow 2006; Shetty 2016).

nHF generates some continuous distending pressure in the airways, which varies with infant weight, flow rate and leak. The pharyngeal pressure generated by nHF is similar to that of nasal CPAP and increases with increased flow (Kubicka 2008; Lampland 2009; Spence 2007; Wilkinson 2008). For a given flow, the delivered pressure appears to be inversely proportional to infant weight (Kubicka 2008; Liew 2020; Wilkinson 2008).

nHF use may provide washout of the nasopharyngeal dead space and subsequent carbon dioxide removal (Dysart 2009). Paediatric (Bressan 2013), animal (Frizzola 2011), and bench top (Sivieri 2017), studies have demonstrated lower carbon dioxide levels in blood with the use of nHF, and one study in preterm neonates also demonstrated flow‐related reductions in nasopharyngeal end‐expiratory carbon dioxide levels (Liew 2020).

The large surface area of the nasopharynx, whilst allowing heating and humidification of inspired gas, causes resistance to inspiratory flow. nHF delivers gas flows above the peak inspiratory flow of the patient, which reduces resistance and work of breathing (Dysart 2009).

The efficacy of non‐invasive respiratory support (CPAP or nHF) in avoiding intubation and mechanical ventilation may be influenced by the approach to treatment with exogenous surfactant. Surfactant therapy improves respiratory mechanics in preterm infants with RDS (Bahadue 2012). Surfactant may be administered to infants receiving CPAP or nHF via brief placement of an endotracheal tube using the INSURE (INtubation, SURfactant, Extubation) procedure (Herting 2013), or via a thin catheter briefly inserted into the trachea (Abdel‐Latif 2021).

Why it is important to do this review

There is currently very little evidence regarding nHF use in extremely preterm infants. With increasing use of nHF, evidence regarding comparative efficacy of the therapy in different gestational age (GA) groups may be of importance to clinicians.

The purpose of this review was to compare the efficacy and harms of nHF with other methods of first‐line non‐invasive respiratory support in preterm infants.

Objectives

To evaluate the benefits and harms of nHF for primary respiratory support in preterm infants compared to other forms of non‐invasive respiratory support.

Methods

Criteria for considering studies for this review

Types of studies

We included all randomised controlled trials (RCTs) and quasi‐randomised studies, including cross‐over studies and cluster‐randomised trials. Trials with a superiority, non‐inferiority or equivalence hypothesis were included. We included studies reported in abstract form in the Characteristics of studies awaiting classification table. For studies with only a subset of relevant participants, we contacted study authors to obtain data.

Types of participants

We included preterm infants (born less than 37 weeks' gestation) receiving non‐invasive respiratory support soon after birth, either prophylactically or for treatment of RDS without a prior period of mechanical ventilation via an endotracheal tube.

The previous versions of this review also included the use of nHF for other indications in preterm infants (Wilkinson 2011; Wilkinson 2016). Due to the increase in number of studies of nHF in preterm infants, we limited this update to studies of nHF as primary respiratory support (see Differences between protocol and review). The efficacy of nHF for postsurfactant or postextubation respiratory support will be assessed in a separate Cochrane Review.

Types of interventions

For the purposes of this review, we defined nHF as the delivery of oxygen or blended oxygen and air via nasal cannulae at gas flows of greater than 1 L/min.

Comparator interventions included:

• nHF compared with CPAP;

• nHF compared with NIPPV;

• nHF compared with ambient oxygen;

• nHF compared with low flow nasal cannulae.

Types of outcome measures

Outcome measures that were not in the previous versions of the review(s) (Wilkinson 2011; Wilkinson 2016; see Differences between protocol and review), or that were modified or included after review of the available data, are marked with an '^a'. We did not exclude studies based on the non‐reporting of outcomes of interest.

Primary outcomes

• Death (before hospital discharge) or BPD (as defined below).

• Death (before hospital discharge).

• BPD, defined as receiving supplemental oxygen or respiratory support (or both) at 36 weeks' postmenstrual age (PMA) for infants born at less than 32 weeks' gestation, or at 28 days of age for infants born at 32 weeks' gestation or greater. Note: data from studies that reported an outcome of 'BPD', 'chronic lung disease' or 'CLD' without a clear accompanying definition were still included in this outcome.

• Treatment failure (as defined in the included studies) within 72 hours of trial entry^a.

• Mechanical ventilation via an endotracheal tube within 72 hours of trial entry^a.

Note: ^a see Types of outcome measures.

Secondary outcomes

Respiratory support

• Mechanical ventilation via an endotracheal tube at any time point following trial entry.

• Duration of mechanical ventilation via an endotracheal tube (days, or PMA at cessation).

• Duration of any form of respiratory support (mechanical ventilation, CPAP, NIPPV, nHF or supplemental oxygen) (days, or PMA at cessation).

• Surfactant administration (via any method).

• Duration of hospitalisation (days, or PMA at hospital discharge).

Complications

• Air leak syndromes (pneumothorax, pneumomediastinum, pneumopericardium or pulmonary interstitial emphysema (PIE)) reported either individually or as a composite outcome (during assigned treatment).

• Nasal trauma (defined as erythema or erosion of the nasal mucosa, nares or septum). Note some studies reported nasal trauma severity as a continuous outcome and could not be included in meta‐analysis (during assigned treatment).

• Nosocomial sepsis (defined as positive blood or cerebrospinal fluid (CSF) cultures taken after five days of age). Note: some studies used alternate definitions, or did not define sepsis: such data were included in the meta‐analysis.

• Gastrointestinal perforation or severe necrotising enterocolitis (NEC) (stage II or more according to Bell's criteria (Bell 1978)). Note: some included studies only reported the incidence of NEC, and were included in the analysis of this outcome.

• Days to attain full feeds^a.

Neurosensory outcomes

• Retinopathy of prematurity (ROP): any stage, and stage 3 or greater.

• Long‐term neurodevelopment (rates of cerebral palsy on physician assessment; developmental delay, i.e. intelligence quotient 2 standard deviations (SD) less than the mean on validated assessment tools such as Bayley's Mental Developmental Index), blindness, hearing impairment requiring amplification.

Note: ^a see Types of outcome measures.

Search methods for identification of studies

For this update, we developed new search strategies to increase sensitivity.

Electronic searches

The Cochrane Neonatal Group Information Specialist, M Fiander, wrote the search strategies in consultation with the review authors. Search strategies are available in Appendix 1.

We searched the following databases without date, language or publication type limits:

• Cochrane Central Register of Controlled Trials via CRS (2022, Issue 3);

• Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process, In‐Data‐Review & Other Non‐Indexed Citations, Daily and Versions (1946 to 10 March 2022);

• CINAHL via EBSCO (1981 to 12 March 2022).

We searched the following clinical trial registries for ongoing or completed trials:

• World Health Organization's International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp/search/en/);

• ISRCTN Registry (www.isrctn.com).

The Cochrane Neonatal Group acknowledges that Embase is a recommended source and will be searched for subsequent updates of this review.

Searching other resources

We searched the reference lists of any articles selected for inclusion in this review in order to identify additional relevant articles.

Data collection and analysis

We used the standard methods of Cochrane Neonatal.

Selection of studies

We included all RCTs and quasi‐randomised trials fulfilling the selection criteria. The review authors (KH, DW, AGDP, BJM) reviewed the results of the search and independently selected the studies for inclusion. The review authors resolved any disagreement by discussion. Search results were managed using Endnote and Covidence software. Selection was recorded in sufficient detail to complete the PRISMA flow diagram (Figure 1).

1.

Study flow diagram.

Data extraction and management

Two review authors (KH and BJM) independently performed trial searches, all review authors performed assessments of methodology and extraction of data, and compared and resolved any differences found at each stage. For each trial, we collected information regarding blinding of randomisation, the intervention and outcome measurements as well as completeness of follow‐up. For any cross‐over trials, only data from the first period were to be included. Where any queries arose or where additional data were required, we made attempts to contact study authors for clarification. For trials in which a review author (i.e. BJM) was an investigator, other review authors performed assessments about eligibility and data extraction for those trials.

Assessment of risk of bias in included studies

Two review authors (KH and BJM) independently assessed the risk of bias (low, high or unclear) of all included trials using the Cochrane RoB 1 tool (Higgins 2011), for the following domains:

• sequence generation (selection bias);

• allocation concealment (selection bias);

• blinding of participants and personnel (performance bias);

• blinding of outcome assessment (detection bias);

• incomplete outcome data (attrition bias);

• selective reporting (reporting bias);

• any other bias.

For trials in which a review author (i.e. BJM) was an investigator, KH and AGDP or DW performed risk of bias assessment for those trials. We resolved any disagreements by discussion or by a third assessor. See Appendix 2 for a more detailed description of risk of bias for each domain.

Measures of treatment effect

We extracted categorical data (e.g. number of deaths or with BPD) for each intervention group, and calculated risk ratio (RR), risk difference (RD) and number needed to treat for an additional beneficial outcome (NNTB), or number needed to treat for an additional harmful outcome (NNTH) as appropriate. We obtained means and SDs for continuous data (e.g. duration of respiratory support, or duration of supplemental oxygen) and calculated mean differences (MD) (where studies used the same scale) or standardised mean difference (SMD) (where studies used different scales). We calculated 95% confidence interval (CI) for each measure of effect.

Unit of analysis issues

The unit of analysis was the participating infant as all trials randomised infants individually, and an infant was considered only once in the analysis.

Dealing with missing data

Where feasible, we carried out analysis on an intention‐to‐treat basis for all outcomes. We requested additional data from study authors where important outcome data were missing, or presented without the GA subgroups of interest (see Subgroup analysis and investigation of heterogeneity). When studies reported only medians for continuous data, and authors were unable to provide means and SDs, we calculated the means using the method described by Wan 2014, as recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020), acknowledging that it was not possible to ascertain whether these outcomes were normally distributed.

Assessment of heterogeneity

We described the clinical diversity and methodological variability of the evidence in the review text and study tables describing study characteristics including design features, population characteristics and intervention details. To assess statistical heterogeneity, we visually inspected forest plots and described the direction and magnitude of effects and the degree of overlap between CIs. We also considered the statistics generated in forest plots that measured statistical heterogeneity. We used the I² statistic to quantify inconsistency amongst the trials in each analysis. We considered the P value from the Chi² test to assess if this heterogeneity was significant (P < 0.1). If we identified substantial heterogeneity, we planned to report the finding and explore possible explanatory factors using prespecified subgroup analysis. Our interpretation of the I² statistic took into account an understanding that measures of heterogeneity will be estimated with high uncertainty when the number of studies is small as recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020).

Assessment of reporting biases

We assessed reporting bias by comparing the stated primary and secondary outcomes and reported outcomes. Where study protocols were available, we compared these to the full publications to determine the likelihood of reporting bias.

Data synthesis

We performed meta‐analysis using Review Manager Web (RevMan Web 2020). For categorical outcomes, we calculated the estimates of RR and RD, each with its 95% CI; for continuous outcomes, we calculated the MD or SMD, each with its 95% CI. We used a fixed‐effect model to combine data where it was reasonable to assume that studies were estimating the same underlying treatment effect. Where there was evidence of clinical heterogeneity, we attempted to explain this based on the different study characteristics and subgroup analyses. If meta‐analysis was not possible, we presented individual trial data separately.

Subgroup analysis and investigation of heterogeneity

We prespecified subgroup analyses for primary outcomes to compare effects by gestation at birth when data were available:

• less than 28 weeks' gestation versus 28 to 32 weeks' gestation versus 32 weeks' gestation or greater.

We planned to explore moderate or high levels of heterogeneity (I² > 50%) in these subgroup analyses of primary outcomes:

• surfactant administration permitted during the intervention period (without this being deemed treatment failure) versus not permitted;

• second‐line CPAP permitted in the nHF arm prior to mechanical ventilation versus not permitted.

See: Differences between protocol and review.

Sensitivity analysis

We planned to perform sensitivity analyses if:

• there was unexplained high heterogeneity (I² > 75%) (explored by removing the outlying trial or trials);

• a trial with high risk of bias (including high level of missing outcome data) was included in the meta‐analysis of an outcome where the other studies had low risk of bias (removed the study with high risk of bias).

Summary of findings and assessment of the certainty of the evidence

We used the GRADE approach, as outlined in the GRADE Handbook to assess the certainty of evidence of the following clinically relevant outcomes (Schünemann 2013):

• death or BPD;

• death;

• BPD;

• treatment failure within 72 hours of trial entry;

• mechanical ventilation within 72 hours of trial entry;

• pneumothorax;

• nasal trauma.

Two review authors (KH and BJM) independently assessed the certainty of the evidence for each of the outcomes above. The trials in which BJM was an investigator were assessed by KH and DW. We considered evidence from RCTs as high certainty but downgraded 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 used GRADEpro GDT to create two summary of findings tables to report the certainty of the evidence (GRADEpro GDT). The summary of findings tables were for two comparisons: nHF compared to CPAP for primary respiratory support in preterm infants (Table 1) and nHF compared to NIPPV for primary respiratory support in preterm infants (Table 2).

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

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

• Moderate certainty: 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.

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

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

Results

Description of studies

We have provided results of the search for this review update in the study flow diagram (Figure 1).

Results of the search

Database searches identified 801 references. After removing 203 duplicates, 598 references were available for title/abstract screening. We excluded 543 records during title/abstract screening and reviewed 55 full‐text articles. We excluded 33 full‐text reports with reasons (Excluded studies; Characteristics of excluded studies table); placed nine potentially relevant studies in awaiting classification (Characteristics of studies awaiting classification table); and included 13 studies (Included studies; Characteristics of included studies table) in our analysis. One included study was published only in abstract form, but we obtained data from the authors and were thus able to include it in this review (Nair 2005). Trial registry searches identified 17 records; of these 13 appear relevant and are ongoing (Characteristics of ongoing studies table; Ongoing studies).

See Figure 1 for details of study selection process.

Several RCTs of nHF compared with other means of non‐invasive support are currently in progress, have been completed but are not yet published or are awaiting further assessment. Nine studies awaiting classification (Awad 2021; Balasubramanian 2022; Cetinkaya 2018; Febre 2015; Iskandar 2019; Lawrence 2012; Oktem 2021; Park 2011; Shirvani 2019) and 13 are ongoing (ACTRN12610000677000; ACTRN12611000233921; CTRI/2017/09/009910; CTRI/2019/10/021633; Irct2016052510026N; Irct20180226038865N; Irct20190623043988N; Irct20200616047788N; ISRCTN66716753; NCT01270581; NCT02055339; NCT02499744; UMIN000018983). See Characteristics of studies awaiting classification and Characteristics of ongoing studies tables for further details.

Included studies

We included 13 studies (2540 preterm infants) in our review (see Characteristics of included studies table).

Of those 13 included studies:

• 11 were available as full journal publications (Armanian 2019; Demirel 2019; Farhat 2018; Kugelman 2015; Lavizzari 2016; Manley 2019; Murki 2018; Roberts 2016; Sharma 2019; Shin 2017; Yoder 2013);

• one was published as an abstract (Nair 2005), but study authors provided additional unpublished data enabling its inclusion in this review;

• Wang 2018 was published in English in abstract form, and in full‐text in Chinese;

• seven study authors kindly provided additional data (Kugelman 2015; Lavizzari 2016; Manley 2019; Murki 2018; Nair 2005; Roberts 2016; Yoder 2013);

• overall, trials enrolled participants from 26 to 36 weeks' gestation (very few infants less than 28 weeks' gestation included in any study): nHF flows ranged from 2.5 L/min to 8 L/min, CPAP pressures ranged from 4 cmH₂O to 8 cmH₂O and NIPPV settings varied (peak inspiratory pressure (PIP) 14 cmH₂O to 22 cmH₂O, PEEP 5 cmH₂O to 6 cmH₂O);

• nine included studies compared nHF with CPAP for primary respiratory support in preterm infants; two compared nHF with NIPPV; and two studies compared nHF with both CPAP and NIPPV;

• no studies compared nHF with ambient oxygen or low flow nasal cannulae;

• where stated, studies were funded internally or by government organisations; some studies had nHF equipment provided by manufacturers.

Comparison 1. Nasal high flow compared with continuous positive airway pressure for primary respiratory support in preterm infants (11 studies)

• Armanian 2019 was a multicentre study in Iran that enrolled 109 preterm neonates with birth weight less than 1500 g (mean GA 30 weeks) and RDS based on clinical examination and chest X‐ray findings. The study was undertaken from February 2015 to March 2016. Infants were randomised to nHF (35 infants; 2.5 L/min for infants less than 1000 g, 3 L/min for infants 1000 g to 1500 g), nasal CPAP (37 infants; pressure 5 cmH₂O to 6 cmH₂O) or NIPPV (37 infants; PIP 16 cmH₂O to 20 cmH₂O, PEEP 5 cmH₂O to 6 cmH₂O, rate 50/min, inspiratory time 0.4 seconds). The primary outcomes were treatment failure (defined by prespecified criteria) and duration of RDS treatment (not defined). Surfactant could be administered by the INSURE technique without this being deemed treatment failure.

• Demirel 2019 was a single‐centre study in Turkey that enrolled 107 preterm infants of 32 weeks' gestation or less who did not require intubation, with or without respiratory distress. The study was undertaken from February 2017 to February 2018. Infants were randomised to nHF (53 infants; initial flow 6 L/min, maximum 8 L/min) or CPAP (54 infants; initial pressure 6 cmH₂O, maximum 7 cmH₂O). The primary outcome was treatment failure, based on prespecified criteria. Surfactant could be administered by the INSURE technique without this being deemed treatment failure.

• Farhat 2018 was a single‐centre study in Iran that enrolled 160 preterm infants 28 to 34 weeks' gestation with respiratory distress. The study was undertaken from April 2014 to November 2014. Infants were randomised to nHF (54 infants; weight‐based, initial flow rate 2 L/min or greater, maximum 5 L/min), CPAP (53 infants; initial pressure 6 cmH₂O, maximum 8 cmH₂O) or NIPPV (53 infants; initial PIP less than 18 cmH₂O, maximum PIP 19 cmH₂O, initial PEEP 4 cmH₂O to 5 cmH₂O, maximum PEEP 6 cmH₂O, rate and inspiratory time not stated). The primary outcomes were need for endotracheal intubation and mechanical ventilation within 72 hours, based on prespecified criteria. Surfactant could be administered by the INSURE technique without this being deemed treatment failure.

• Lavizzari 2016 was a single‐centre non‐inferiority study in Italy that enrolled 316 preterm infants 29 to 36 weeks' gestation with mild–moderate respiratory distress. The study was undertaken from 5 January 2012 to 28 June 2014. Infants were randomised to nHF (158 infants; initial flow 4 L/min to 6 L/min, maximum 6 L/min) or nasal CPAP (158 infants; initial pressure 4 cmH₂O to 6 cmH₂O, maximum 6 cmH₂O). The primary outcome was need for mechanical ventilation within 72 hours of commencing respiratory support, based on prespecified criteria. The use of bilevel positive airway pressure (BiPAP) was permitted in the CPAP arm without this being deemed treatment failure. Surfactant was administered via the INSURE technique if the fraction of inspired oxygen (FiO₂) increased to greater than 0.35, without this being deemed treatment failure.

• Manley 2019 was a multicentre non‐inferiority study in non‐tertiary special care nurseries in Australia that enrolled 754 term and preterm infants. Of these, 379 infants were preterm and were included in this review. The study was undertaken from 13 April 2015 to 28 November 2017. Infants were less than 24 hours of age, GA 31 weeks or greater and birth weight greater than 1200 g with respiratory distress requiring non‐invasive respiratory support or supplemental oxygen. Infants were randomised to nHF (185 preterm infants; initial gas flow 6 L/min, maximum 8 L/min) or CPAP (194 preterm infants; initial pressure 6 cmH₂O, maximum 8 cmH₂O). The primary outcome was treatment failure, defined by prespecified criteria. Infants in the nHF group who reached treatment failure criteria could receive CPAP with the aim of avoiding mechanical ventilation. Surfactant was only administered following endotracheal intubation and mechanical ventilation.

• Murki 2018 was a multicentre non‐inferiority study in India that enrolled 272 preterm infants 28 to 36 weeks' gestation with respiratory distress. The study was undertaken from 9 October 2015 to 26 November 2016. Infants were randomised to nHF (133 infants; initial flow 5 L/min, maximum 7 L/min) or CPAP (139 infants; initial pressure 5 cmH₂O, maximum 7 cmH₂O). The primary outcome was treatment failure, defined by prespecified criteria. Infants in the nHF group who reached treatment failure criteria could receive CPAP with the aim of avoiding mechanical ventilation. Surfactant could be administered by the INSURE technique without this being deemed treatment failure.

• Nair 2005 was a single‐centre study in the USA, published in abstract form, that enrolled 67 preterm infants of 27 to 34 weeks' gestation with respiratory distress in the first six hours after birth. Infants were randomised to nHF (33 infants; initial flow 5 L/min to 6 L/min, maximum not stated) or CPAP (34 infants; initial pressure 5 cmH₂O to 6 cmH₂O, maximum not stated). The primary outcome was respiratory failure requiring endotracheal intubation, based on prespecified criteria. After randomisation, surfactant was only administered following endotracheal intubation and mechanical ventilation.

• Roberts 2016 was a multicentre non‐inferiority study in Australia and Norway that enrolled 564 preterm infants 28 to 36 weeks' gestation with respiratory distress. The study was undertaken from 27 May 2013 to 16 June 2015. Infants were randomised to nHF (289 infants; initial gas flow 6 L/min to 8 L/min, maximum 8 L/min) or CPAP (294 infants; initial pressure 6 cmH₂O to 8 cmH₂O, maximum 8 cmH₂O). The primary outcome was treatment failure within 72 hours of randomisation, defined by prespecified criteria. Infants in the nHF group who reached treatment failure criteria could receive second‐line CPAP with the aim of avoiding mechanical ventilation. Surfactant was only administered following endotracheal intubation and mechanical ventilation.

• Sharma 2019 was a single‐centre study in India that enrolled 100 infants 26 to 34+6 weeks' gestation with mild–moderate respiratory distress within six hours of birth. Infants were randomised to nHF (50 infants) or CPAP (50 infants). There was no information regarding set flow rate or pressure. The primary outcomes were duration of non‐invasive support, duration of oxygen supplementation and treatment failure (no prespecified criteria). The approach to surfactant treatment was not stated.

• Shin 2017 was a single‐centre non‐inferiority study in Korea that enrolled 87 preterm infants 30 to 35 weeks' gestation less than 24 hours old with respiratory distress. The study was undertaken from August 2010 to August 2013. Infants were randomised to nHF (42 infants; initial flow rate 5 L/min, maximum 7 L/min) or CPAP (43 infants; initial pressure 5 cmH₂O, maximum 7 cmH₂O). The primary outcome was the incidence of treatment failure (intubation and mechanical ventilation, time frame not defined), based on prespecified criteria. After treatment failure criteria were met, infants in the nHF group could receive CPAP, and infants in the CPAP group could receive BiPAP, with the aim of avoiding mechanical ventilation. Surfactant was only administered following endotracheal intubation and mechanical ventilation.

• Yoder 2013 was a multicentre study in the USA and China that enrolled 432 term and preterm infants of more than 28 weeks' gestation at birth who were planned to receive non‐invasive respiratory support either as primary support after birth or postextubation. Of all infants enrolled, 351 infants were preterm, with 125 preterm infants in the primary support arm that were included in this review. The study was undertaken from December 2007 to April 2012. Infants were randomised to nHF (58 preterm infants; initial flow 3 L/min to 5 L/min, maximum 6 L/min to 8 L/min) or CPAP (67 preterm infants; initial pressure 5 cmH₂O to 6 cmH₂O, maximum 8 cmH₂O). The primary outcome was need for intubation within 72 hours of commencing the allocated treatment, based on prespecified criteria. After randomisation, surfactant was only administered following endotracheal intubation and mechanical ventilation.

Comparison 2. Nasal high flow compared with nasal intermittent positive pressure ventilation for primary respiratory support in preterm infants (four studies)

• Armanian 2019 was a multicentre study in Iran that enrolled 109 preterm neonates (mean GA 30 weeks) with respiratory distress to nHF, CPAP or NIPPV (see details under Comparison 1).

• Farhat 2018 was a single‐centre study in Iran that enrolled 160 preterm infants 28 to 34 weeks' gestation with respiratory distress to nHF, CPAP or NIPPV (see details under Comparison 1).

• Kugelman 2015 was a single‐centre study in Iran that enrolled 76 preterm infants born less than 35 weeks' gestation, with birth weight greater than 1000 g, who required non‐invasive respiratory support. Infants were treated with either nHF (38 infants; initial flow 1 L/min, maximum 5 L/min) or synchronised NIPPV (38 infants; PIP 14 cmH₂O to 22 cmH₂O, PEEP 6 cmH₂O, rate 12/min to 30/min). The primary outcome was treatment failure defined by prespecified criteria. Following randomisation, surfactant was only administered following endotracheal intubation and mechanical ventilation.

• Wang 2018 was a single‐centre study in China that enrolled 89 preterm infants 28 to 32 weeks' gestation and birth weight 1000 g to 1500 g with respiratory distress. Infants were randomised to nHF (43 infants; initial flow 5 L/min, maximum 8 L/min) or NIPPV (46 infants; initial PIP 18 cmH₂O, initial PEEP 6 cmH₂O, maximum not stated). The primary outcome was need for intubation and mechanical ventilation within 72 hours after birth, based on prespecified criteria. Surfactant was only administered following endotracheal intubation and mechanical ventilation.

Comparison 3. Nasal high flow compared with ambient oxygen

No studies compared nHF with ambient oxygen.

Comparison 4. Nasal high flow compared with low flow nasal cannulae

No studies compared nHF with low flow nasal cannulae.

Excluded studies

We excluded 12 new studies in this update.

In total we excluded 33 studies for the reasons described in the Characteristics of excluded studies table.

Studies awaiting classification

There are nine studies awaiting classification (Awad 2021; Balasubramanian 2022; Cetinkaya 2018; Febre 2015; Iskandar 2019; Lawrence 2012; Oktem 2021; Park 2011; Shirvani 2019). See Characteristics of studies awaiting classification table for further details.

Ongoing studies

There are 13 are ongoing studies (ACTRN12610000677000; ACTRN12611000233921; CTRI/2017/09/009910; CTRI/2019/10/021633; Irct2016052510026N; Irct20180226038865N; Irct20190623043988N; Irct20200616047788N; ISRCTN66716753; NCT01270581; NCT02055339; NCT02499744; UMIN000018983). See Characteristics of ongoing studies table for further details.

Risk of bias in included studies

See Figure 2; Figure 3.

2.

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

3.

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

Allocation

Six studies described appropriate random sequence generation and allocation concealment (Demirel 2019; Lavizzari 2016; Manley 2019; Murki 2018; Roberts 2016; Shin 2017). Armanian 2019 and Sharma 2019 reported incomplete information regarding the process of allocation concealment. Farhat 2018 described a 'randomised convenience sampling method' that matched infants on birth weight and GA, which indicates a potential for selection bias due to lack of both random sequence generation and allocation concealment. Kugelman 2015 and Wang 2018 did not fully describe the methods of random sequence generation.

Blinding

Blinding of treatment allocation was not attempted in any study except Sharma 2019, where authors stated the study was double‐blinded, but gave no information regarding the process of blinding. Due to the nature of the intervention, blinding was unlikely in these studies.

There were prespecified criteria for treatment failure and intubation in all studies except Sharma 2019. However, in several studies, some criteria were potentially open to bias (e.g. duration of FiO₂ increase, frequency of blood gas analysis, and recording of apnoea frequency and severity). Lack of blinding was also a potential source of bias for subjective outcomes such as the presence of nasal mucosal injury.

In Kugelman 2015, alterations to flow or the level of non‐invasive support were at the discretion of treating clinicians, rather than based on objective criteria. Four studies did not state which criteria they used for escalating non‐invasive support (Armanian 2019; Farhat 2018; Nair 2005; Sharma 2019).

Four trials permitted the use of CPAP in infants who met nHF treatment failure criteria, after the primary outcome was determined, prior to endotracheal intubation and mechanical ventilation (Manley 2019; Murki 2018; Roberts 2016; Shin 2017). In these four studies, a number of infants in the nHF group meeting treatment failure criteria did not subsequently require mechanical ventilation. Two studies permitted the use of BiPAP in infants who met CPAP failure criteria prior to endotracheal intubation and mechanical ventilation (Lavizzari 2016; Shin 2017). Five studies permitted the administration of surfactant via INSURE prior to meeting treatment failure criteria, in both infants receiving nHF and CPAP or NIPPV (Armanian 2019; Demirel 2019; Farhat 2018; Lavizzari 2016; Murki 2018). While all trials had objective failure criteria which were to be met prior to switching to CPAP or BiPAP, or administering surfactant, there may still be potential for bias between groups due to clinician preference.

Incomplete outcome data

Three studies did not report the number of eligible infants, the number who received the allocated intervention or whether the analysis was performed using an intention‐to‐treat principle (Demirel 2019; Farhat 2018; Wang 2018). Wang 2018 excluded an unknown number of infants who were discharged prior to completing the allocated intervention. Shin 2017 excluded one infant from each arm of the trial prior to analysis. Sharma 2019 did not perform an intention‐to‐treat analysis; infants who underwent mechanical ventilation were excluded from outcome reporting.

Selective reporting

Trial registration was not evident, or occurred after trial completion for seven studies, indicating the potential for selective reporting of outcomes (Demirel 2019; Farhat 2018; Lavizzari 2016; Nair 2005; Sharma 2019; Shin 2017; Wang 2018). The trial registration by Armanian 2019 only included two of the three intervention arms present in the published study. Sharma 2019 did not report data for several listed outcomes, and Wang 2018 excluded infants who were discharged from the analysis. Lavizzari 2016 had previously reported interim outcome data from their trial prior to achieving the planned sample size (Ciuffini 2014).

Other potential sources of bias

Three studies reported continuous outcomes such as durations of respiratory support, supplemental oxygen and hospitalisation as means and SD (Murki 2018; Nair 2005; Yoder 2013). Farhat 2018 reported means but no SDs. The remaining studies reported medians and interquartile ranges (IQRs) for continuous outcomes, and these were converted to means and SD using the method described by Wan 2014, and recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020). These outcome data may not have been normally distributed.

Effects of interventions

See: Table 1; Table 2

Comparison 1. Nasal high flow compared with continuous positive airway pressure for primary respiratory support in preterm infants

See Table 1.

We included 11 studies (total 2196 infants) in this comparison (Armanian 2019; Demirel 2019; Farhat 2018; Lavizzari 2016; Manley 2019; Nair 2005; Murki 2018; Roberts 2016; Sharma 2019; Shin 2017; Yoder 2013). Shin 2017 did not report a time frame for the outcomes of treatment failure and mechanical ventilation, therefore, we did not include the data; however, these findings were consistent with the results of the meta‐analysis. Nair 2005 reported both treatment failure and mechanical ventilation at seven days, not 72 hours, therefore, the data were not included; in this study, equal numbers of infants in each arm required mechanical ventilation at seven days.

Primary outcomes

Death or bronchopulmonary dysplasia (Analysis 1.1)

Meta‐analysis of data from seven studies (1830 infants) suggests that the use of nHF compared with CPAP may result in little to no difference in death or BPD (Analysis 1.1).

1.1. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 1: Death or bronchopulmonary dysplasia

• RR 1.09, 95% CI 0.74 to 1.60 (I² = 0%)

• RD 0, 95% CI −0.02 to 0.02

We found no evidence of subgroup differences by gestation (Chi² = 1.62, degrees of freedom (df) = 2 (P = 0.45), I² = 0%).

Subgroup analyses for heterogeneity

Not applicable (I² = 0%).

Sensitivity analyses

Risk of bias: not applicable (I² = 0%).

Certainty of evidence (GRADE)

We assessed the certainty of evidence as low, downgraded two levels for very serious imprecision (Table 1).

Death (Analysis 1.2)

Meta‐analysis of data from nine studies (2009 infants) suggests that use of nHF compared with CPAP may result in little to no difference in death (Analysis 1.2).

1.2. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 2: Death

• RR 0.78, 95% CI 0.44 to 1.39 (I² = 0%)

• RD −0.01, 95% CI −0.02 to 0.01

We found no evidence of subgroup differences by gestation (Chi² = 2.03, df = 2 (P = 0.36), I² = 1.3%).

Subgroup analyses for heterogeneity

Not applicable (I² = 1.3%).

Sensitivity analyses

Risk of bias: not applicable (I² = 1.3%).

Certainty of evidence (GRADE)

We assessed the certainty of evidence as low, downgraded two levels for very serious imprecision (Table 1).

Bronchopulmonary dysplasia (Analysis 1.3)

Meta‐analysis of data from eight studies (1917 infants) suggests that use of nHF compared with CPAP may result in little to no difference in BPD (Analysis 1.3).

1.3. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 3: Bronchopulmonary dysplasia

• RR 1.14, 95% CI 0.74 to 1.76 (I² = 0%)

• RD 0, 95% CI −0.01 to 0.02

We found no evidence of subgroup differences by gestation (Chi² = 1.11, df = 2 (P = 0.57), I² = 0%).

Subgroup analyses for heterogeneity

Not applicable (I² = 0%).

Sensitivity analyses

Risk of bias: not applicable (I² = 0%).

Certainty of evidence (GRADE)

We assessed the certainty of evidence as low, downgraded two levels for very serious imprecision (Table 1).

Treatment failure (as defined in the included studies) within 72 hours of trial entry (Analysis 1.4)

Meta‐analysis of data from nine studies (2042 infants) suggests that use of nHF compared with CPAP likely results in an increase in treatment failure within 72 hours of trial entry (Analysis 1.4).

1.4. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 4: Treatment failure within 72 hours of trial entry

• RR 1.70, 95% CI 1.41 to 2.06 (I² = 36%)

• RD 0.09, 95% CI 0.06 to 0.12

• NNTH 11, 95% CI 8 to 17

On subgroup analysis by GA, there was no difference between the subgroups (test for subgroup differences comparing 28 to 32 weeks' GA and 32 weeks' GA or greater: P = 0.58, I² = 0%).

Subgroup analyses for heterogeneity

Not applicable (I² = 0%).

Sensitivity analyses

Risk of bias: not applicable (I² = 0%).

Certainty of evidence (GRADE)

We assessed the certainty of evidence as moderate, downgraded one level for serious risk of bias (Table 1).

Nair 2005 only reported treatment failure at seven days, and Shin 2017 did not report a time frame for treatment failure, therefore we excluded the data from these studies. Five studies provided data for GA subgroups (Lavizzari 2016; Manley 2019; Murki 2018; Roberts 2016; Yoder 2013). There were very few infants of less than 28 weeks' GA included in any study.

Mechanical ventilation via an endotracheal tube within 72 hours of trial entry (Analysis 1.5)

Meta‐analysis of data from nine studies (2042 infants) suggests that the use of nHF compared with CPAP likely does not increase mechanical ventilation within 72 hours of trial entry (Analysis 1.5).

1.5. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 5: Mechanical ventilation within 72 hours of trial entry

• RR 1.04, 95% CI 0.82 to 1.31

• RD 0.00, 95% CI −0.02 to 0.03

We found no evidence of subgroup differences by gestation (Chi² = 3.32, df = 2 (P = 0.19), I² = 39.7%).

Subgroup analyses for heterogeneity

Not applicable (I² = 39.7%).

Sensitivity analyses

Risk of bias: not applicable (I² = 39.7%).

Certainty of evidence (GRADE)

We assessed the certainty of evidence as moderate, downgraded one level for serious risk of bias (Table 1).

On subgroup analysis, there was no difference in the risk of mechanical ventilation when trials that permitted surfactant use via INSURE during the intervention period were excluded (RR 1.19, 95% CI 0.86 to 1.66; 3 studies, 1068 infants; Analysis 1.17). There was no difference in the risk of mechanical ventilation when trials that permitted second‐line CPAP during the intervention period were excluded (RR 0.86, 95% 0.62 to 1.20; 6 studies, 827 infants; Analysis 1.19). Five studies provided data for GA subgroups (Lavizzari 2016; Manley 2019; Murki 2018; Roberts 2016; Yoder 2013).

1.17. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 17: Subgroup analysis – mechanical ventilation with or without surfactant permitted

1.19. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 19: Subgroup analysis – mechanical ventilation with or without second‐line CPAP permitted

Secondary outcomes

The use of nHF as primary respiratory support, compared with CPAP, likely results in a reduction in nasal trauma (RR 0.49, 95% CI 0.36 to 0.68; RD −0.06, 95% CI −0.09 to −0.04; 7 studies, 1595 infants; moderate‐certainty evidence; Analysis 1.12). Infants managed with nHF had a longer duration of respiratory support (MD 0.52 days, 95% CI 0.25 to 0.80; 7 studies, 1808 infants; Analysis 1.7). nHF probably reduces pneumothorax compared with CPAP (RR 0.66, 95% CI 0.40 to 1.08; 10 studies, 2094 infants; moderate‐certainty evidence; Analysis 1.11). A subgroup analysis excluding trials which permitted surfactant use via INSURE during the intervention period indicated that, without surfactant availability, nHF use was associated with a reduced risk of pneumothorax, compared with CPAP (RR 0.48, 95% CI 0.23 to 1.00; RD −0.02, 95% CI −0.04 to 0.00; 5 studies, 1220 infants; Analysis 1.18). Other secondary outcomes were similar between groups, including mechanical ventilation at any time point after trial entry (Analysis 1.6), duration of oxygen supplementation (Analysis 1.8), surfactant treatment (Analysis 1.9), duration of hospitalisation (Analysis 1.10), nosocomial sepsis (Analysis 1.13), gastrointestinal perforation or severe NEC (Analysis 1.14), time to full feeds (Analysis 1.15), and ROP (Analysis 1.16). No studies reported other individual air leak syndromes or long‐term neurodevelopmental outcomes.

1.12. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 12: Nasal trauma

1.7. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 7: Duration of any respiratory support (days)

1.11. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 11: Pneumothorax

1.18. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 18: Subgroup analysis – pneumothorax with or without surfactant permitted

1.6. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 6: Mechanical ventilation at any time point after trial entry

1.8. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 8: Duration of supplemental oxygen (days)

1.9. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 9: Surfactant treatment

1.10. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 10: Duration of hospitalisation (days)

1.13. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 13: Nosocomial sepsis

1.14. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 14: Gastrointestinal perforation or severe necrotising enterocolitis

1.15. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 15: Time to full feeds (days)

1.16. Analysis.

Comparison 1: Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants, Outcome 16: Retinopathy of prematurity

Comparison 2. Nasal high flow compared with nasal intermittent positive pressure ventilation for primary respiratory support in preterm infants

See Table 2.

We included four studies (total 343 infants) in this comparison (Armanian 2019; Farhat 2018; Kugelman 2015; Wang 2018).

Primary outcomes

Death or bronchopulmonary dysplasia (Analysis 2.1)

Meta‐analysis of data from two studies (total 182 infants) suggests that the use of nHF compared with NIPPV may have little to no effect on death or BPD but the evidence is very uncertain (Analysis 2.1).

2.1. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 1: Death or bronchopulmonary dysplasia

• RR 0.64, 95% CI 0.30 to 1.37

• RD −0.05, 95% CI −0.14 to 0.04

We found no evidence of subgroup differences by gestation (Chi² = 2.92, df = 3 (P = 0.40), I² = 0%).

Subgroup analyses for heterogeneity

Not applicable (I² = 0%).

Sensitivity analyses

Risk of bias: not applicable (I² = 0%).

Certainty of evidence (GRADE)

We assessed the certainty of evidence as very low, downgraded two levels for very serious imprecision and one level for serious risk of bias (Table 2).

Death (Analysis 2.2)

Meta‐analysis of data from three studies (254 infants) suggests that the use of nHF compared with NIPPV may result in little to no difference in death (Analysis 2.2).

2.2. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 2: Death

• RR 0.78, 95% CI 0.36 to 1.69

• RD −0.02, 95% CI −0.10 to 0.05

We found no evidence of subgroup differences by gestation (Chi² = 0.27, df = 3 (P = 0.97), I² = 0%).

Subgroup analyses for heterogeneity

Not applicable (I² = 0%).

Sensitivity analyses

Risk of bias: not applicable (I² = 0%).

Certainty of evidence (GRADE)

We assessed the certainty of evidence as low, downgraded two levels for very serious imprecision (Table 2).

Bronchopulmonary dysplasia (Analysis 2.3)

Meta‐analysis of data from three studies (271 infants) suggests that the use of nHF compared with NIPPV may result in little to no difference in BPD (Analysis 2.3).

2.3. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 3: Bronchopulmonary dysplasia

• RR 1.19, 95% CI 0.66 to 2.12

• RD 0.02, 95% CI −0.05 to 0.10

We found no evidence of subgroup differences by gestation (Chi² = 1.65, df = 3 (P = 0.65), I² = 0%).

Subgroup analyses for heterogeneity

Not applicable (I² = 0%).

Sensitivity analyses

Risk of bias: not applicable (I² = 0%).

Certainty of evidence (GRADE)

We assessed the certainty of evidence as low, downgraded two levels for very serious imprecision (Table 2).

Treatment failure (Analysis 2.4)

Meta‐analysis of data from four studies (343 infants) suggests that the use of nHF compared with NIPPV likely results in little to no difference in treatment failure within 72 hours of trial entry (Analysis 2.4).

2.4. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 4: Treatment failure within 72 hours of trial entry

• RR 1.27, 95% CI 0.90 to 1.79

• RD 0.06, 95% CI −0.03 to 0.15

We found no evidence of subgroup differences by gestation (Chi² = 1.65, df = 3 (P = 0.65), I² = 0%).

Subgroup analyses for heterogeneity

Not applicable (I² = 0%).

Sensitivity analyses

Risk of bias: not applicable (I² = 0%).

Certainty of evidence (GRADE)

We assessed the certainty of evidence as moderate, downgraded one level for serious risk of bias (Table 2).

Mechanical ventilation (Analysis 2.5)

Meta‐analysis of data from four studies (343 infants) suggests that the use of nHF compared with NIPPV likely results in little to no difference in mechanical ventilation within 72 hours of trial entry (Analysis 2.5).

2.5. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 5: Mechanical ventilation within 72 hours of trial entry

• RR 0.91, 95% CI 0.62 to 1.33

• RD −0.02, 95% CI −0.11 to 0.06

We found no evidence of subgroup differences by gestation (Chi² = 0.09, df = 3 (P = 0.99), I² = 0%).

Subgroup analyses for heterogeneity

Not applicable (I² = 0%).

Sensitivity analyses

Risk of bias: not applicable (I² = 0%).

Certainty of evidence (GRADE)

We assessed the certainty of evidence as moderate, downgraded one level for serious risk of bias (Table 2).

Data on any primary outcome for GA subgroups were available in Kugelman 2015 and Wang 2018. Kugelman 2015 was the only study to enrol extremely preterm infants born less than 28 weeks' GA; however, there were only three such infants included.

Secondary outcomes

The use of nHF as primary respiratory support, compared with NIPPV, likely results in a reduction in nasal trauma (RR 0.21, 95% CI 0.09 to 0.47; RD −0.17, 95% CI −0.24 to −0.10; 3 studies, 272 infants; moderate‐certainty evidence) (Analysis 2.10). There were no differences in other secondary outcomes between groups, including mechanical ventilation at any time after trial entry (Analysis 2.6), surfactant treatment (Analysis 2.7), duration of hospitalisation (Analysis 2.8), pneumothorax (Analysis 2.9), nosocomial sepsis (Analysis 2.11), gastrointestinal perforation or severe NEC (Analysis 2.12), or time to full feeds (Analysis 2.13). When study authors reported only medians for continuous data, and were unable to provide means (SD), we had planned to calculate the means using the method described by Wan 2014, as recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020). However, only one study reported duration of respiratory support (Analysis 2.14) or duration of supplemental oxygen (Analysis 2.15); it was not possible to ascertain whether these outcomes were normally distributed and therefore accurately able to be converted to means. Kugelman 2015 reported duration of respiratory support to be a median of four days (IQR 1 to 15) in the nHF group compared to two days (IQR 0.3 to 6.5) in the NIPPV group. Wang 2018 reported duration of supplemental oxygen as a median of 13.7 days (IQR 4.9 to 29) in the nHF group, compared to 12.6 days (IQR 5.4 to 25.8) in the NIPPV group.

2.10. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 10: Nasal trauma

2.6. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 6: Mechanical ventilation at any time point after trial entry

2.7. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 7: Surfactant treatment

2.8. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 8: Duration of hospitalisation (days)

2.9. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 9: Pneumothorax

2.11. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 11: Nosocomial sepsis

2.12. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 12: Gastrointestinal perforation or severe necrotising enterocolitis

2.13. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 13: Time to full feeds (days)

2.14. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 14: Duration of respiratory support (days)

Duration of respiratory support (days)
Study	Nasal HF – median (IQR)	NIPPV – median (IQR)
Kugelman 2015	4 (1‐15)	2 (0.3‐6.5)

2.15. Analysis.

Comparison 2: Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants, Outcome 15: Duration of supplemental oxygen (days)

Duration of supplemental oxygen (days)
Study	Nasal HF – median (IQR)	NIPPV – median (IQR)
Wang 2018	13.7 (4.9‐29)	12.6 (5.4‐25.8)

No studies reported long‐term neurodevelopmental outcomes.

Comparison 3. Nasal high flow compared with ambient oxygen

No studies examined this comparison.

Comparison 4. Nasal high flow compared with low‐flow nasal cannulae

No studies examined this comparison.

Discussion

Summary of main results

We included 13 RCTs (2540 preterm infants) that compared nHF with other forms of non‐invasive respiratory support for primary respiratory support. Nine studies compared nHF with CPAP for primary respiratory support, two studies compared nHF with NIPPV and two studies compared nHF with both CPAP and NIPPV. Studies varied in the GA of included infants, the nHF gas flows used and the devices used. They also differed regarding the approach to the use of surfactant, and the use of other respiratory support modalities in the event of treatment failure.

When compared with CPAP for primary respiratory support in preterm infants, nHF may result in little to no difference in the rate of death or BPD (low‐certainty evidence). We found that compared with CPAP, the use of nHF likely results in an increase in treatment failure (moderate‐certainty evidence), but not of endotracheal intubation and mechanical ventilation within 72 hours of trial entry (moderate‐certainty evidence), with no difference between the GA subgroups. Subgroup analyses showed similar rates of mechanical ventilation between groups after excluding trials in which surfactant was permitted, and in which second‐line CPAP was permitted.

nHF likely results in a reduction in nasal trauma compared with CPAP (moderate‐certainty evidence). nHF was associated with an approximately half day longer duration of respiratory support; the clinical significance of this is uncertain and may depend on local resources. The rate of pneumothorax was likely lower with the use of nHF, compared with CPAP (moderate‐certainty evidence). In trials where surfactant use was not permitted during the intervention period, the rate of pneumothorax was also lower in babies managed with nHF. Other secondary outcomes including durations of supplemental oxygen and hospitalisation and sepsis were similar between groups.

When compared with NIPPV for primary respiratory support in preterm infants, nHF may have little to no effect on death or BPD but the evidence is very uncertain. nHF likely results in no difference in treatment failure or mechanical ventilation within 72 hours of trial entry, compared with NIPPV. nHF likely results in a reduction in nasal trauma, but other secondary outcomes were similar between groups.

Overall completeness and applicability of evidence

The 13 included studies varied in quality. None of the studies were blinded; while all studies except Sharma 2019 had prespecified criteria for treatment failure and reintubation, the lack of blinding may have led to some bias. We did not perform a sensitivity analysis for study quality.

Four trials permitted the use of CPAP in infants who met nHF treatment failure criteria (Manley 2019; Murki 2018; Roberts 2016; Shin 2017). Two studies also permitted the use of BiPAP in infants who met CPAP failure criteria (Lavizzari 2016; Shin 2017). A subgroup analysis excluding trials in which 'rescue' CPAP was permitted in the event of nHF treatment failure found similar rates of mechanical ventilation.

Furthermore, the approach to surfactant therapy differed between studies. Five studies permitted the use of surfactant via the INtubation, SURfactant, Extubation (INSURE) technique without this being deemed treatment failure (Armanian 2019; Demirel 2019; Farhat 2018; Lavizzari 2016; Murki 2018). The success or failure of non‐invasive respiratory support may differ in infants with atelectatic, surfactant‐deplete lungs, compared with surfactant‐replete lungs.

Overall, most studies reported the primary outcomes of this review. No studies reported long‐term neurodevelopmental outcomes, such as cerebral palsy or developmental delay. Very few extremely preterm infants less than 28 weeks' gestation were included in any study and therefore conclusions cannot be drawn regarding the use of nHF in this population.

Quality of the evidence

The overall certainty of evidence comparing nHF with CPAP was rated using GRADE as low to moderate. The outcomes of treatment failure, mechanical ventilation and nasal trauma were downgraded due to serious concerns regarding risk of bias (lack of blinding of the intervention arms). The outcomes of death and BPD were downgraded due to very serious concerns regarding imprecision, due to the low event rate and wide CIs. The outcome of pneumothorax was also downgraded due to serious concerns regarding imprecision, relating to the low event rate.

Potential biases in the review process

We made the judgement to divide the previous review of nHF in preterm infants into two, given the increase in the number of studies. This current review outlines the evidence for nHF as primary respiratory support (prior to mechanical ventilation or surfactant provision via an endotracheal tube). Studies were included in this review if the main intent of the trial was the provision of nHF for primary respiratory support; we resolved any uncertainties by consensus agreement if required. Iranpour 2012 was included in the primary support section of the original review (Wilkinson 2011), and the 2016 update (Wilkinson 2016), but not this version following communication from the authors suggesting that the randomisation occurred after surfactant provision. We amended the time frame for the outcomes of treatment failure and mechanical ventilation for this review. This was a decision made following selection of the studies, given that all but one study which reported the outcomes (Nair 2005) did so at 72 hours, rather than seven days (the time frame in the previous version of the review). The exclusion of the data from Nair 2005 is a potential bias; we have therefore described the findings for these outcomes in the text.

The protocol and earlier versions of this review included Embase as a search source. This update omitted Embase based on the rationale that CENTRAL now includes records from Embase. This rationale has, subsequently, been flagged as a method that may reduce sensitivity of the search. Subsequent updates will include Embase as a source.

Agreements and disagreements with other studies or reviews

The results of our review are in agreement with a rapid systematic review (Conte 2018) and systematic review (Bruet 2021), and in contrast to the conclusions of the previous version of this review (Wilkinson 2016), given the large number of different included studies of high flow for primary respiratory support.

Authors' conclusions

Implications for practice.

We found that the use of nasal high flow (nHF) for primary respiratory support in preterm infants of 28 weeks' gestation or greater may result in little to no difference in the clinically important outcomes of death or bronchopulmonary dysplasia (BPD), compared with continuous positive airway pressure (CPAP) or nasal intermittent positive pressure ventilation (NIPPV). The overall certainty of evidence for the outcomes of death and BPD was very low to low. The use of nHF likely results in an increase in treatment failure within 72 hours of trial entry, but no difference in the rate of endotracheal intubation and mechanical ventilation, where there is the option of surfactant or CPAP (or both) in the event of nHF treatment failure.

Compared with CPAP, nHF likely results in a reduction in nasal trauma and likely a reduction in pneumothorax. There may be a small increase in the total duration of respiratory support in infants treated with nHF. There is insufficient evidence to guide the use of nHF in extremely preterm infants of less than 28 weeks' gestation based on this review.

Implications for research.

There are nine studies awaiting classification and 13 ongoing studies. Future trials should focus on investigating novel uses of nHF for neonates, including using higher gas flows than current standard practice, in combination with less‐invasive surfactant administration, during neonatal endotracheal intubation, and for delivery room stabilisation of preterm infants. Few of the current ongoing studies will investigate these areas. There is a need for studies evaluating the use of nHF as primary support in extremely preterm infants born less than 28 weeks' gestation and also comparing commercially available devices.

Standardised definitions of nHF flows and treatment failure (with predefined criteria for hypoxia, hypercarbia and apnoea) should be employed in future studies.

What's new

Date	Event	Description
5 May 2023	New citation required and conclusions have changed	Nasal HF is associated with a higher rate of treatment failure, but not mechanical ventilation, compared with CPAP. There is little evidence for the use of nHF as primary respiratory support in extremely preterm infants.
5 May 2023	New search has been performed	We updated the search in March 2022, and identified the 13 studies (total of 2540 preterm infants) included in this review. Methods altered to include only studies of nasal high flow for primary respiratory support (not postextubation or postsurfactant support).

History

Protocol first published: Issue 1, 2007
Review first published: Issue 5, 2011

Date	Event	Description
1 March 2016	New search has been performed	This updates the review "High flow nasal cannula for respiratory support in preterm infants". (Wilkinson 2011).
1 March 2016	New citation required and conclusions have changed	Updated search January 2016.
14 February 2012	Amended	Correction to denominator in Comparison 2. Figures reordered.

Appendices

Appendix 1. Search strategies 2022

MEDLINE strategy

Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process, In‐Data‐Review & Other Non‐Indexed Citations, Daily and Versions <1946 to March 11, 2022>

1 (high flow or high frequency or high gas flow or high gas flows).mp. (101096)

2 (nasal or binasal).mp. (143205)

3 exp Nasal Cavity/ (12366)

4 exp Nasal Mucosa/ (27389)

5 2 or 3 or 4 (150606)

6 (cannula* or prong*).mp. (52383)

7 exp Catheters/ or exp Cannula/ or exp Catheterization/ (223676)

8 6 or 7 (264567)

9 1 and 5 and 8 (2021)

10 (Optiflow or Vapotherm or (Fisher adj2 Paykel)).mp. (193)

11 (hfnc or hfnp or hhfnox or HHHFNC).mp. (937)

12 9 or 10 or 11 (2192)

13 exp infant, newborn/ (648472)

14 (newborn* or new born or new borns or newly born or baby* or babies or premature or prematurity or preterm or pre term or low birth weight or low birthweight or VLBW or LBW or infant or infants or "infant s" or infant's or infantile or infancy or neonat*).ti,ab. (912282)

15 13 or 14 (1203411)

16 randomized controlled trial.pt. (561036)

17 controlled clinical trial.pt. (94734)

18 randomized.ab. (553471)

19 placebo.ab. (226281)

20 drug therapy.fs. (2455314)

21 randomly.ab. (377656)

22 trial.ab. (590648)

23 groups.ab. (2321557)

24 or/16‐23 (5285012)

25 exp animals/ not humans.sh. (4970671)

26 24 not 25 [RCT Filter] (4598867)

27 15 and 26 (206591)

28 12 and 27 (219)

CINAHL (Ebsco)

Search date: March 12, 2022

S1

( (("high flow" OR "high frequency" OR "high gas flow" OR "high gas flows") AND (nasal OR binasal) AND (cannula* OR prong*)) OR ((Optiflow OR Vapotherm OR (Fisher AND Paykel))) OR (hfnc OR hfnp OR hhfnox OR HHHFNC)) ) AND ( (infant or infants or infantís or infantile or infancy or newborn* or "new born" or "new borns" or "newly born" or neonat* or baby* or babies or premature or prematures or prematurity or preterm or preterms or "pre term" or premies or "low birth weight" or "low birthweight" or VLBW or LBW) AND (randomized controlled trial OR controlled clinical trial OR randomized OR randomised OR placebo OR clinical trials as topic OR randomly OR trial OR PT clinical trial) )

136

Cochrane CENTRAL via CRS

Search date: 12 March 2022

	Cochrane CENTRAL via CRS
	12‐Mar‐22
1	high flow or high frequency or high gas flow or high gas flows AND CENTRAL:TARGET	8584
2	nasal or binasal AND CENTRAL:TARGET	21415
3	MESH DESCRIPTOR Nasal Cavity EXPLODE ALL AND CENTRAL:TARGET	428
4	MESH DESCRIPTOR Nasal Mucosa EXPLODE ALL AND CENTRAL:TARGET	1003
5	#2 OR #3 OR #4	21455
6	cannula* or prong* AND CENTRAL:TARGET	6609
7	MESH DESCRIPTOR Catheters EXPLODE ALL AND CENTRAL:TARGET	1983
8	MESH DESCRIPTOR Cannula EXPLODE ALL AND CENTRAL:TARGET	164
9	MESH DESCRIPTOR Catheterization EXPLODE ALL AND CENTRAL:TARGET	9846
10	#6 OR #7 OR #8 OR #9	16564
11	#5 AND #10 AND #1	1265
12	(Optiflow or Vapotherm or (Fisher ADJ2 Paykel)) AND CENTRAL:TARGET	300
13	hfnc or hfnp or hhfnox or HHHFNC AND CENTRAL:TARGET	666
14	#11 OR #12 OR #13	1502
15	MESH DESCRIPTOR Infant, Newborn EXPLODE ALL AND CENTRAL:TARGET	17489
16	infant or infants or infant's or "infant s" or infantile or infancy or newborn* or "new born" or "new borns" or "newly born" or neonat* or baby* or babies or premature or prematures or prematurity or preterm or preterms or "pre term" or premies or "low birth weight" or "low birthweight" or VLBW or LBW or ELBW or NICU AND CENTRAL:TARGET	96314
17	#15 OR #16	96314
18	#14 AND #17	446

Appendix 2. Risk of bias tool

We used the Cochrane RoB 1 tool (Higgins 2011).

Sequence generation (checking for possible selection bias). Was the allocation sequence adequately generated?

For each included study, we categorised the method used to generate the allocation sequence as:

• low risk (any truly random process, e.g. random number table; computer random number generator);

• high risk (any non‐random process, e.g. odd or even date of birth; hospital or clinic record number); or

• unclear risk.

Allocation concealment (checking for possible selection bias). Was allocation adequately concealed?

For each included study, we categorised the method used to conceal the allocation sequence as:

• low risk (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);

• high risk (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth); or

• unclear risk.

Blinding of participants and personnel (checking for possible performance bias). Was knowledge of the allocated intervention adequately prevented during the study?

For each included study, we categorised the methods used to blind study participants and personnel from knowledge of which intervention a participant received. Blinding was assessed separately for different outcomes or class of outcomes. We categorised the methods as:

• low risk, high risk or unclear risk for participants; and

• low risk, high risk or unclear risk for personnel.

Blinding of outcome assessment (checking for possible detection bias). Was knowledge of the allocated intervention adequately prevented at the time of outcome assessment?

For each included study, we categorised the methods used to blind outcome assessment. Blinding was assessed separately for different outcomes or class of outcomes. We categorised the methods as:

• low risk for outcome assessors;

• high risk for outcome assessors; or

• unclear risk for outcome assessors.

Incomplete outcome data (checking for possible attrition bias through withdrawals, dropouts, protocol deviations). Were incomplete outcome data adequately addressed?

For each included study and for each outcome, we described the completeness of data including attrition and exclusions from the analysis. We noted whether attrition and exclusions were reported, the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and whether missing data were balanced across groups or were related to outcomes. Where sufficient information was reported or supplied by the trial authors, we re‐included missing data in the analyses. We categorised the methods as:

• low risk (less than 20% missing data);

• high risk (20% or greater missing data); or

• unclear risk.

Selective reporting bias. Are reports of the study free of suggestion of selective outcome reporting?

For each included study, we described how we investigated the possibility of selective outcome reporting bias and what we found. For studies in which study protocols were published in advance, we compared prespecified outcomes versus outcomes eventually reported in the published results. If the study protocol was not published in advance, we contacted study authors to gain access to the study protocol. We assessed the methods as:

• low risk (where it is clear that all of the study's prespecified outcomes and all expected outcomes of interest to the review have been reported);

• high risk (where not all the study's prespecified outcomes have been reported; one or more reported primary outcomes were not prespecified outcomes of interest and are reported incompletely and so cannot be used; study fails to include results of a key outcome that would have been expected to have been reported); or

• unclear risk.

Other sources of bias. Was the study apparently free of other problems that could put it at a high risk of bias?

For each included study, we described any important concerns we had about other possible sources of bias (for example, whether there was a potential source of bias related to the specific study design or whether the trial was stopped early due to some data‐dependent process). We assessed whether each study was free of other problems that could put it at risk of bias as:

• low risk;

• high risk; or

• unclear risk.

If needed, we explored the impact of the level of bias through undertaking sensitivity analyses.

Data and analyses

Comparison 1. Nasal high flow (nHF) compared with continuous positive airway pressure (CPAP) for primary respiratory support in preterm infants.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Death or bronchopulmonary dysplasia	7	1830	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.74, 1.60]
1.1.1 28–32 weeks	5	567	Risk Ratio (M‐H, Fixed, 95% CI)	1.17 [0.72, 1.89]
1.1.2 ≥ 32 weeks	5	1089	Risk Ratio (M‐H, Fixed, 95% CI)	1.48 [0.58, 3.80]
1.1.3 < 37 weeks (subgroup data not available)	2	174	Risk Ratio (M‐H, Fixed, 95% CI)	0.61 [0.24, 1.53]
1.2 Death	9	2009	Risk Ratio (M‐H, Fixed, 95% CI)	0.78 [0.44, 1.39]
1.2.1 28–32 weeks	5	567	Risk Ratio (M‐H, Fixed, 95% CI)	1.43 [0.36, 5.64]
1.2.2 ≥ 32 weeks	5	1089	Risk Ratio (M‐H, Fixed, 95% CI)	1.40 [0.29, 6.85]
1.2.3 < 37 weeks' (subgroup data not available)	4	353	Risk Ratio (M‐H, Fixed, 95% CI)	0.57 [0.28, 1.16]
1.3 Bronchopulmonary dysplasia	8	1917	Risk Ratio (M‐H, Fixed, 95% CI)	1.14 [0.74, 1.76]
1.3.1 28–32 weeks	5	567	Risk Ratio (M‐H, Fixed, 95% CI)	1.13 [0.68, 1.88]
1.3.2 ≥ 32 weeks	5	1089	Risk Ratio (M‐H, Fixed, 95% CI)	1.70 [0.57, 5.04]
1.3.3 < 37 weeks (subgroup data not available)	3	261	Risk Ratio (M‐H, Fixed, 95% CI)	0.56 [0.12, 2.58]
1.4 Treatment failure within 72 hours of trial entry	9	2042	Risk Ratio (M‐H, Fixed, 95% CI)	1.70 [1.41, 2.06]
1.4.1 28–32 weeks	5	567	Risk Ratio (M‐H, Fixed, 95% CI)	2.11 [1.50, 2.95]
1.4.2 ≥ 32 weeks	5	1089	Risk Ratio (M‐H, Fixed, 95% CI)	1.85 [1.36, 2.52]
1.4.3 < 37 weeks (subgroup data not available)	4	386	Risk Ratio (M‐H, Fixed, 95% CI)	1.12 [0.78, 1.60]
1.5 Mechanical ventilation within 72 hours of trial entry	9	2042	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.82, 1.31]
1.5.1 28–32 weeks	5	567	Risk Ratio (M‐H, Fixed, 95% CI)	1.33 [0.90, 1.97]
1.5.2 ≥ 32 weeks	5	1089	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.69, 1.52]
1.5.3 < 37 weeks (subgroup data not available)	4	386	Risk Ratio (M‐H, Fixed, 95% CI)	0.78 [0.51, 1.19]
1.6 Mechanical ventilation at any time point after trial entry	4	1175	Risk Ratio (M‐H, Fixed, 95% CI)	1.19 [0.92, 1.55]
1.7 Duration of any respiratory support (days)	7	1808	Mean Difference (IV, Fixed, 95% CI)	0.52 [0.25, 0.80]
1.8 Duration of supplemental oxygen (days)	6	1723	Mean Difference (IV, Fixed, 95% CI)	‐0.07 [‐0.20, 0.05]
1.9 Surfactant treatment	8	1590	Risk Ratio (M‐H, Fixed, 95% CI)	0.99 [0.87, 1.13]
1.10 Duration of hospitalisation (days)	7	1808	Mean Difference (IV, Fixed, 95% CI)	‐0.16 [‐1.54, 1.21]
1.11 Pneumothorax	10	2094	Risk Ratio (M‐H, Fixed, 95% CI)	0.66 [0.40, 1.08]
1.12 Nasal trauma	7	1595	Risk Ratio (M‐H, Fixed, 95% CI)	0.49 [0.36, 0.68]
1.13 Nosocomial sepsis	9	2022	Risk Difference (M‐H, Fixed, 95% CI)	‐0.01 [‐0.03, 0.01]
1.14 Gastrointestinal perforation or severe necrotising enterocolitis	6	1469	Risk Difference (M‐H, Fixed, 95% CI)	0.00 [‐0.01, 0.01]
1.15 Time to full feeds (days)	6	1741	Mean Difference (IV, Fixed, 95% CI)	‐0.27 [‐0.76, 0.22]
1.16 Retinopathy of prematurity	4	1259	Risk Difference (M‐H, Fixed, 95% CI)	0.00 [‐0.01, 0.01]
1.17 Subgroup analysis – mechanical ventilation with or without surfactant permitted	8	1942	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.81, 1.30]
1.17.1 Surfactant permitted	5	874	Risk Ratio (M‐H, Fixed, 95% CI)	0.86 [0.61, 1.21]
1.17.2 Surfactant not permitted	3	1068	Risk Ratio (M‐H, Fixed, 95% CI)	1.19 [0.86, 1.66]
1.18 Subgroup analysis – pneumothorax with or without surfactant permitted	10	2094	Risk Ratio (M‐H, Fixed, 95% CI)	0.66 [0.40, 1.08]
1.18.1 Surfactant not permitted	5	1220	Risk Ratio (M‐H, Fixed, 95% CI)	0.48 [0.23, 1.00]
1.18.2 Surfactant permitted	5	874	Risk Ratio (M‐H, Fixed, 95% CI)	0.89 [0.45, 1.76]
1.19 Subgroup analysis – mechanical ventilation with or without second‐line CPAP permitted	9	2042	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.80, 1.28]
1.19.1 Rescue CPAP permitted	3	1215	Risk Ratio (M‐H, Fixed, 95% CI)	1.17 [0.85, 1.63]
1.19.2 Rescue CPAP not permitted	6	827	Risk Ratio (M‐H, Fixed, 95% CI)	0.86 [0.62, 1.20]

Comparison 2. Nasal high flow (nHF) compared with nasal intermittent positive pressure ventilation (NIPPV) for primary respiratory support in preterm infants.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Death or bronchopulmonary dysplasia	2	182	Risk Difference (M‐H, Fixed, 95% CI)	‐0.05 [‐0.14, 0.04]
2.1.1 < 28 weeks	1	3	Risk Difference (M‐H, Fixed, 95% CI)	0.50 [‐0.32, 1.32]
2.1.2 28–32 weeks	1	28	Risk Difference (M‐H, Fixed, 95% CI)	‐0.02 [‐0.22, 0.18]
2.1.3 ≥ 32 weeks	1	44	Risk Difference (M‐H, Fixed, 95% CI)	0.00 [‐0.08, 0.08]
2.1.4 < 37 weeks (no subgroup data available)	1	107	Risk Difference (M‐H, Fixed, 95% CI)	‐0.10 [‐0.23, 0.04]
2.2 Death	3	254	Risk Difference (M‐H, Fixed, 95% CI)	‐0.02 [‐0.10, 0.05]
2.2.1 < 28 weeks	1	3	Risk Difference (M‐H, Fixed, 95% CI)	0.00 [‐0.73, 0.73]
2.2.2 28–32 weeks	1	28	Risk Difference (M‐H, Fixed, 95% CI)	0.00 [‐0.13, 0.13]
2.2.3 ≥ 32 weeks	1	44	Risk Difference (M‐H, Fixed, 95% CI)	0.00 [‐0.08, 0.08]
2.2.4 < 37 weeks (subgroup data not available)	2	179	Risk Difference (M‐H, Fixed, 95% CI)	‐0.03 [‐0.13, 0.07]
2.3 Bronchopulmonary dysplasia	3	271	Risk Ratio (M‐H, Fixed, 95% CI)	1.19 [0.66, 2.12]
2.3.1 < 28 weeks	1	3	Risk Ratio (M‐H, Fixed, 95% CI)	1.50 [0.38, 6.00]
2.3.2 28–32 weeks	2	117	Risk Ratio (M‐H, Fixed, 95% CI)	1.23 [0.61, 2.48]
2.3.3 ≥ 32 weeks	1	0	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
2.3.4 < 37 weeks (no subgroup data available)	1	107	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.26, 3.72]
2.4 Treatment failure within 72 hours of trial entry	4	343	Risk Ratio (M‐H, Fixed, 95% CI)	1.27 [0.90, 1.79]
2.4.1 < 28 weeks	1	3	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.39, 2.58]
2.4.2 28–32 weeks	2	117	Risk Ratio (M‐H, Fixed, 95% CI)	0.97 [0.45, 2.09]
2.4.3 > 32 weeks	1	44	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.33, 2.55]
2.4.4 < 37 weeks (no subgroup data available)	2	179	Risk Ratio (M‐H, Fixed, 95% CI)	1.50 [0.96, 2.33]
2.5 Mechanical ventilation within 72 hours of trial entry	4	343	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.62, 1.33]
2.5.1 < 28 weeks	1	3	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.39, 2.58]
2.5.2 28–32 weeks	2	117	Risk Ratio (M‐H, Fixed, 95% CI)	0.97 [0.45, 2.09]
2.5.3 ≥ 32 weeks	1	44	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.33, 2.55]
2.5.4 < 37 weeks (no subgroup data available)	2	179	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.52, 1.47]
2.6 Mechanical ventilation at any time point after trial entry	2	183	Risk Ratio (M‐H, Fixed, 95% CI)	0.81 [0.56, 1.17]
2.7 Surfactant treatment	4	344	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.71, 1.15]
2.8 Duration of hospitalisation (days)	1	76	Mean Difference (IV, Fixed, 95% CI)	1.00 [‐9.10, 11.10]
2.9 Pneumothorax	4	344	Risk Ratio (M‐H, Fixed, 95% CI)	0.78 [0.40, 1.53]
2.10 Nasal trauma	3	272	Risk Difference (M‐H, Fixed, 95% CI)	‐0.17 [‐0.24, ‐0.10]
2.11 Nosocomial sepsis	2	183	Risk Ratio (M‐H, Fixed, 95% CI)	0.69 [0.28, 1.74]
2.12 Gastrointestinal perforation or severe necrotising enterocolitis	2	165	Risk Ratio (M‐H, Fixed, 95% CI)	1.88 [0.41, 8.65]
2.13 Time to full feeds (days)	1	76	Mean Difference (IV, Fixed, 95% CI)	1.70 [‐1.21, 4.61]
2.14 Duration of respiratory support (days)	0		Other data	No numeric data
2.15 Duration of supplemental oxygen (days)	0		Other data	No numeric data

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Armanian 2019.

Study characteristics
Methods	Randomised controlled trial
Participants	109 preterm neonates with BW < 1500 g; mean 30 weeks' GA
Interventions	nHF (35 infants): 2.5 L/min for infants < 1000 g, 3 L/min for infants 1000–1500 g Nasal CPAP (37 infants): pressure 5–6 cmH2O NIPPV (37 infants): termed NIMV in this study, PIP 16–20 cmH2O, PEEP 5–6 cmH2O, rate 50 breaths/min, inspiratory time 0.4 seconds
Outcomes	Treatment failure (based on prespecified criteria: pH < 7.2 and PCO2 > 60 mmHg, repeated or severe apnoea, FiO2 > 40–60%, or requirement for emergency intubation). Surfactant Duration of respiratory support IVH PDA Pneumothorax Duration of supplemental oxygen Duration of hospitalisation Mortality
Notes	Recruitment from February 2015 to March 2016 at 2 centres in Iran. No funding stated, no conflicts of interest declared. Infants in nHF group had a higher GA (approx 1 week) and higher BW (approx 130 g). Infants received surfactant via INSURE if FiO2 > 0.3 to maintain SpO2 > 91%. Registered trial protocol did not have CPAP group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Permuted block randomisation (block size 6).
Allocation concealment (selection bias)	Unclear risk	Envelopes containing names of therapy group (not stated whether opaque, sequentially numbered and sealed).
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Treatment failure criteria stated; however, large range for FiO2 and duration not stated, blood gas criteria unclear.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Similar reasons for missing secondary outcome data across groups.
Selective reporting (reporting bias)	High risk	Trial registry did not have CPAP arm. Not all outcomes from trial registry reported in paper.
Other bias	Low risk	No other bias noted.

Demirel 2019.

Study characteristics
Methods	Randomised controlled trial
Participants	107 preterm infants ≤ 32 weeks' gestation with spontaneous respiration
Interventions	nHF (54 infants): Vapotherm device, initial flow rate 6 L/min, up to a maximum of 8 L/min Nasal CPAP (53 infants): SERVO‐i ventilator, initial pressure 6 cmH2O, up to a maximum of 8 cmH2O
Outcomes	Primary outcome Treatment failure, based on prespecified criteria: increase in work of breathing; apnoea/bradycardia/desaturation > 3 times in 1 hour; apnoea requiring positive pressure ventilation FiO2 > 0.6 to maintain SpO2 90% to 95% pH < 7.25 or pCO2 > 65 mmHg Secondary outcomes Duration of non‐invasive respiratory support Duration of oxygen therapy Maximum FiO2 Duration of hospital stay Intubation Pneumothorax BPD
Notes	Recruitment from February 2017 to February 2018 at 1 centre in Turkey. No funding stated, no conflicts of interest declared. All eligible infants ≤ 32 weeks' GA enrolled (did not need to have respiratory distress). Surfactant given via INSURE if FiO2 > 0.4 and RDS on CXR.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random number.
Allocation concealment (selection bias)	Low risk	Sequentially numbered sealed opaque envelopes containing group assignments.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Caregivers unblinded to intervention. Treatment failure criteria specified; however, included a subjective assessment of increased 'work of breathing' and apnoea/bradycardia/desaturation.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Insufficient reporting of attrition/exclusions to permit judgement.
Selective reporting (reporting bias)	Unclear risk	Unclear whether trial was registered prospectively.
Other bias	Low risk	No other bias noted.

Farhat 2018.

Study characteristics
Methods	Randomised controlled trial
Participants	160 preterm neonates 28–34+6 weeks' gestation; BW 800–2500 g with RDS
Interventions	nHF therapy (54 infants): flow rate dependent upon weight (2–5 L/min) Nasal CPAP (53 infants): initial pressure 6 cmH2O, maximum pressure 8 cmH2O Nasal intermittent positive pressure ventilation (53 infants): initial PIP < 18 cmH2O, initial PEEP 4–5 cmH2O, rate and inspiratory time not stated
Outcomes	Primary outcome Intubation and mechanical ventilation within 72 hours, based on prespecified criteria: severe apnoea PaCO2 > 60 mmHg pH < 7.2 FiO2 > 0.5 Secondary outcomes Duration of mechanical ventilation Duration of oxygen therapy Surfactant administration IVH Sepsis Air leak Pulmonary haemorrhage Nasal trauma BPD Death
Notes	Recruitment from April 2014 to November 2014 at 1 centre in Iran. No funding stated, no conflicts of interest declared. Randomisation and consent processes not outlined. Unclear starting flow rates for nHF and uncertain which devices used. Surfactant permitted via INSURE. No data for the combined outcome of death or BPD.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Stated to be 'randomised convenience sampling method.' Information about sequence generation not available.
Allocation concealment (selection bias)	High risk	No description in paper.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Criteria stated; however, 'severe apnoea' not defined.
Blinding of outcome assessment (detection bias) Nasal trauma	High risk	No description of assessment in paper.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Insufficient reporting of attrition/exclusions to permit judgement.
Selective reporting (reporting bias)	Unclear risk	Trial not prospectively registered.
Other bias	Low risk	No other bias noted.

Kugelman 2015.

Study characteristics
Methods	Randomised controlled trial
Participants	76 preterm infants < 35 weeks' gestation, BW > 1000 g requiring primary respiratory support from birth
Interventions	HFNC (38 infants): 1–5 L/min Synchronised NIPPV via nasal prongs (38 infants): 12–30 cycles/min, inspiratory time of 0.3 seconds, PEEP 6 cmH2O, and PIP 14–22 cmH2O
Outcomes	Treatment failure (increased respiratory distress plus respiratory acidosis (pH < 7.2, CO2 > 60 mmHg) FiO2 > 0.50 or recurrent significant apnoea) Duration of respiratory support Clinical features IVH BPD (oxygen at 36 weeks' PMA) Time until full feeds Length of stay Air leak Nasal trauma Gastrointestinal perforation
Notes	Recruitment at 1 centre in Israel. Pilot study, underpowered to detect clinically significant difference in outcomes between interventions. Time point for treatment failure not defined. Infants meeting failure of treatment criteria in the HFNC arm were able to receive NIPPV. Treatment failure reported separately from reintubation. nHF equipment supplied by Vapotherm. No external funding, no conflicts of interest declared.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	'Randomly prepared cards.' Process of sequence generation not described.
Allocation concealment (selection bias)	Low risk	Sealed envelopes.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Unblinded study. Criteria stated.
Blinding of outcome assessment (detection bias) Nasal trauma	High risk	Unblinded assessment. Criteria not stated in main paper or protocol.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data.
Selective reporting (reporting bias)	Low risk	Trial registered. All outcomes recorded in protocol reported.
Other bias	Low risk	No other bias noted.

Lavizzari 2016.

Study characteristics
Methods	Randomised non‐inferiority trial
Participants	316 preterm infants 29–36+6 weeks' GA with mild–moderate respiratory distress
Interventions	nHF (158 infants): Vapotherm Precision Flow, initial flow rate 4–6 L/min, maximum flow rate 6 L/min Nasal CPAP (158 infants): SiPAP, Viasys Healthcare, initial pressure 4–6 cmH2O, maximum pressure 6 cmH2O
Outcomes	Primary outcome Mechanical ventilation (excluding INSURE) within 72 hours, based on prespecified criteria: FiO2 > 0.4 to maintain SpO2 86–93% > 4 episodes apnoea per hour or > 2 requiring positive pressure ventilation pH < 7.20 PaCO2 > 70 mmHg Secondary outcomes Duration of respiratory support Surfactant treatment Days to reach full enteral feeds Duration of hospitalisation Pneumothorax IVH PDA Sepsis NEC BPD ROP Death
Notes	Recruitment from 5 January 2012 to 28 June 2014 at 1 centre in Italy. No funding stated, no conflicts of interest declared. Surfactant permitted prior to reaching treatment failure criteria, administered via INSURE if FiO2 > 0.35 to maintain SpO2 86–93%. BiPAP also permitted in CPAP arm if apnoea or increased work of breathing. Trial not registered prior to commencement. Preliminary publication of some results prior to trial completion. Further outcome data for GA subgroups provided by study authors.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomisation.
Allocation concealment (selection bias)	Low risk	Sequentially numbered, sealed, opaque envelopes.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Unblinded study with prespecified criteria for treatment failure; however, some subjectivity, e.g. definition of 'persistent' oxygen requirement or respiratory acidosis.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data.
Selective reporting (reporting bias)	Unclear risk	Trial only registered after completion of recruitment. Preliminary publication of results prior to trial completion.
Other bias	Unclear risk	Preliminary publication of results (Ciuffini 2014).

Manley 2019.

Study characteristics
Methods	Randomised non‐inferiority trial
Participants	754 infants (379 preterm infants 31–36+6 weeks' GA and BW > 1200 g) with respiratory distress in non‐tertiary special care units
Interventions	nHF (185 infants): Fisher & Paykel Optiflow Junior, initial gas flow 6 L/min, maximum gas flow 8 L/min Nasal CPAP (194 infants): binasal prongs or nasal mask: initial pressure 6 cmH2O, maximum pressure 8 cmH2O
Outcomes	Primary outcome Treatment failure within 72 hours, based on prespecified criteria: maximal support plus FiO2 ≥ 0.4 pH ≤ 7.2 pCO2 > 60 mmHg ≤ 2 episodes of apnoea requiring positive pressure ventilation within 24 hours or 6 episodes requiring any intervention within 6 hours urgent need for intubation or NICU transfer as determined by treating clinician Secondary outcomes Reasons for treatment failure Intubation Transfer to tertiary NICU Surfactant Duration of respiratory support and supplemental oxygen Hospital stay Death before hospital discharge BPD Pneumothorax Cost of care
Notes	Recruitment from 13 April 2015 to 28 November 2017 at 9 non‐tertiary special‐care nurseries in Australia. Funded by Australian National Health and Medical Research Council and Monash University. No conflicts of interest declared in publication. Lead author is also an author on this Cochrane Review (BJM). Infants assigned to nHF arm who met treatment failure criteria could receive CPAP as a rescue therapy. Surfactant not permitted prior to intubation. Further outcome data for GA subgroups provided by study authors.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation sequence with variable block sizes.
Allocation concealment (selection bias)	Low risk	Sequentially numbered, sealed, opaque envelopes.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Unblinded study. Treatment failure criteria prespecified; however, some subjectivity, e.g. included urgent need for intubation as per treating clinician.
Blinding of outcome assessment (detection bias) Nasal trauma	Unclear risk	Assessor not blinded to treatment group, but objective criteria for assessment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data.
Selective reporting (reporting bias)	Low risk	Trial registered and all outcomes reported.
Other bias	Low risk	No other bias noted.

Murki 2018.

Study characteristics
Methods	Randomised non‐inferiority trial
Participants	272 preterm infants 28–36+6 weeks' GA and BW > 1000 g with respiratory distress in 2 centres in India
Interventions	nHF (133 infants): Fisher & Paykel Optiflow Junior or AIRVO 2, initial flow rate 5 L/min, maximum 7 L/min Nasal CPAP (139 infants): Fisher & Paykel, binasal prongs or mask, initial pressure 5 cmH2O, maximum pressure 7 cmH2O
Outcomes	Primary outcome Treatment failure within 72 hours of life, based on prespecified criteria: maximal support and FiO2 > 0.6 pH < 7.20 PaCO2 > 60 mmHg increase in Silverman Anderson score recurrent apnoea (4 per hour or any requiring positive pressure ventilation) shock Secondary outcomes Mechanical ventilation within 72 hours of life Mechanical ventilation within 7 days of life Surfactant Mortality Duration of respiratory support Duration of oxygen therapy Pneumothorax Culture‐positive sepsis NEC Bell Stage ≥ II Nasal injury BPD Grade III/IV IVH Cystic PVL ROP Time to full feeds Duration of hospitalisation Weight at discharge
Notes	Recruitment from 9 October 2015 to 26 November 2016 at 2 centres in India. No funding stated, no conflicts of interest disclosed. Rescue CPAP permitted in nHF group following treatment failure, prior to intubation. Surfactant permitted via INSURE if RDS on CXR and FiO2 > 0.3. Incorrect sample size calculation. Trial registration stated different sample size to final paper. Further outcome data for GA subgroups provided by study authors.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation.
Allocation concealment (selection bias)	Low risk	Serially numbered, opaque, sealed envelopes.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Unblinded study, prespecified treatment failure criteria but some subjectivity with interpretation (e.g. duration of increased FiO2).
Blinding of outcome assessment (detection bias) Nasal trauma	Unclear risk	Assessor not blinded to intervention arm but objective scoring method.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data.
Selective reporting (reporting bias)	Unclear risk	Trial prospectively registered but discrepancies between registration and final protocol.
Other bias	Low risk	No other bias noted.

Nair 2005.

Study characteristics
Methods	Randomised controlled trial
Participants	67 preterm infants with respiratory distress requiring CPAP in 1st 6 hours, 27–34 weeks' GA (mean 32 weeks)
Interventions	HFNC (33 infants): Vapotherm 5–6 L/min CPAP (34 infants): bubble CPAP, Hudson prongs, 5–6 cmH2O
Outcomes	Respiratory failure (leading to intubation) (pH ≤ 7.25 and PaCO2 ≥ 60 mmHg, or FiO2 > 0.70, or severe or frequent apnoea) Nasal injury BPD (as defined in Jobe 2001) Mortality Length of hospitalisation Sepsis Pneumothorax
Notes	Recruitment at 1 centre in the USA. Study finished prior to achieving target sample size due to recall of Vapotherm units. Study authors provided a full‐text manuscript including results. Vapotherm provided equipment for the study. No information regarding conflicts of interest available.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Stratified into 27–30 weeks' and 31–34 weeks' GA. Permuted block randomisation.
Allocation concealment (selection bias)	Low risk	Sealed envelopes.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Standardised criteria for respiratory failure, though frequency of blood gases and recording of apnoea not blinded.
Blinding of outcome assessment (detection bias) Nasal trauma	High risk	Assessment of nasal injury non‐blinded. Assessment method not stated.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Insufficient reporting of attrition/exclusions to permit judgement.
Selective reporting (reporting bias)	Unclear risk	Trial not registered.
Other bias	Low risk	No other bias noted.

Roberts 2016.

Study characteristics
Methods	Randomised non‐inferiority trial
Participants	564 preterm infants 28–36+6 weeks' GA with early respiratory distress
Interventions	nHF (289 infants): initial gas flow 6–8 L/min, either Fisher & Paykel Optiflow Junior or Vapotherm Precision Flow Nasal CPAP (285 infants): initial pressure 6–8 cmH2O, by ventilator, bubble or variable‐flow
Outcomes	Primary outcome Treatment failure within 72 hours, based on prespecified criteria: maximal support plus FiO2 ≥ 0.4 pH ≤ 7.2 and pCO2 > 60 mmHg ≥ 2 episodes of apnoea requiring positive pressure ventilation within 24 hours or 6 episodes requiring any intervention within 6 hours urgent need for intubation as determined by treating clinician Secondary outcomes Mechanical ventilation within 72 hours Nasal trauma Other complications of prematurity Pneumothorax Death Cost of care
Notes	Recruitment from 27 May 2013 to 16 June 2015 at 9 centres in Australia and Norway. Funded by the Australian National Health and Medical Research Council. 1 study author declared potential of conflict of interest due to travel support from Fisher & Paykel. Another study author is an author on this Cochrane Review (BJM). Cross‐over to CPAP group allowed in event of nHF failure, prior to intubation. Surfactant not permitted prior to intubation. Further outcome data for GA subgroups provided by study authors.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation sequence with variable block sizes.
Allocation concealment (selection bias)	Low risk	Sequentially numbered, sealed, opaque envelopes opened when eligibility and consent criteria met.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Unblinded study, but objective treatment failure criteria. Reasons for treatment failure reported, and treatment failure reported separately to intubation.
Blinding of outcome assessment (detection bias) Nasal trauma	Unclear risk	Assessor unblinded to treatment arm but used validated scoring chart.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data.
Selective reporting (reporting bias)	Low risk	Trial registered and protocol published.
Other bias	Low risk	No other bias noted.

Sharma 2019.

Study characteristics
Methods	Randomised controlled trial
Participants	100 infants 26–34+6 weeks' GA with mild–moderate respiratory distress within 6 hours of birth
Interventions	nHF (50 infants): no information regarding device or flow rates Nasal CPAP (50 infants): no information regarding device or pressure
Outcomes	Primary outcomes Duration of non‐invasive support Duration of oxygen supplementation Treatment failure (no criteria for definition) Secondary outcomes Time to full feeds Nasal trauma Air leak BPD PDA ROP
Notes	Recruitment at 1 centre in India. No funding stated, no conflicts of interest declared. Paper stated study was double‐blind but no information given regarding process of blinding. No information regarding device, flow rate or pressure. Not intention‐to‐treat analysis (infants who required mechanical ventilation excluded from analysis). Several secondary outcomes not reported. No trial registration.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random numbers.
Allocation concealment (selection bias)	High risk	No information regarding allocation concealment.
Blinding of participants and personnel (performance bias)	Unclear risk	Paper stated "double‐blinded" but no information regarding this process.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	High risk	No objective criteria for intubation. Paper stated "double‐blinded" but no information regarding this process.
Blinding of outcome assessment (detection bias) Nasal trauma	High risk	Assessment method not stated.
Incomplete outcome data (attrition bias) All outcomes	High risk	Reason for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups.
Selective reporting (reporting bias)	High risk	Trial not registered. No data reported for several secondary outcomes. Not intention‐to‐treat reporting of primary outcomes.
Other bias	Low risk	No other bias noted.

Shin 2017.

Study characteristics
Methods	Randomised non‐inferiority trial
Participants	87 preterm infants 30–35 weeks' GA aged < 24 hours with respiratory distress
Interventions	nHF (42 infants): Fisher & Paykel Optiflow, initial flow rate 5 L/min, maximum 7 L/min Nasal CPAP (43 infants): Infant Flow, initial pressure 5 cmH2O, maximum 7 cmH2O
Outcomes	Primary outcome Treatment failure (time point not defined), based on prespecified criteria of maximal support and: pH < 7.2 and PaCO2 > 65 mmHg FiO2 > 0.4 apnoea (> 2–3 episodes/hour) Secondary outcomes Duration of respiratory support Air leak Nasal trauma Surfactant BPD (NIH consensus definition) Symptomatic PDA (requiring pharmacological or surgical treatment) IVH ≥ grade III PVL; sepsis (defined as bacteraemia) NEC ≥ stage 2 (modified Bell's) Days to full enteral feeds
Notes	Recruitment from August 2010 to August 2013 at 1 centre in Korea. No funding stated, no conflicts of interest disclosed. 20% non‐inferiority margin. Infants in nHF arm could receive CPAP in event of treatment failure; infants in CPAP arm could receive BiPAP.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random numbers.
Allocation concealment (selection bias)	Low risk	Sealed opaque envelopes, opened once consent/inclusion criteria met.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Non‐blinded study, but treatment failure prespecified (some subjectivity, e.g. range for apnoea).
Blinding of outcome assessment (detection bias) Nasal trauma	High risk	Assessment method not stated.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data.
Selective reporting (reporting bias)	Unclear risk	Not strictly intention‐to‐treat analysis (1 infant excluded from analysis due to later diagnosis with congenital heart disease, 1 due to wrong device applied).
Other bias	Low risk	No other bias noted.

Wang 2018.

Study characteristics
Methods	Randomised trial
Participants	89 preterm infants 28–32+6 weeks' GA and BW 1000–1500 g with respiratory distress
Interventions	nHF (43 infants): Fisher & Paykel Optiflow Junior, initial flow rate 5 L/min Nasal intermittent positive pressure ventilation (46 infants): SLE5000 Ventilator, initial PIP 18 cmH2O, PEEP 6 cmH2O, rate 40 breaths/min
Outcomes	Primary outcome Treatment failure (endotracheal intubation and surfactant administration) within 72 hours, based on prespecified criteria: pH < 7.2 and PaCO2 > 60 mmHg PaO2 < 50 mmHg FiO2 > 0.4 NEC apnoea (> 6 episodes in 6 hours, ≥ 2 episodes requiring PPV) Secondary outcomes Surfactant therapy Pneumonia Duration of invasive ventilation Duration of non‐invasive ventilation BPD Moderate‐to‐severe BPD NEC ≥ Stage IIB ROP Grade III–IV IVH PDA Nasal injury Air leak
Notes	Recruitment at 1 centre in China. No funding stated, no information regarding conflicts of interest available. Full text in Chinese, translated to English by Dr Wei Qi Fan. No cross‐over or 'rescue' CPAP use. No surfactant permitted prior to intubation. No information on consent or age at randomisation. No sample size calculation or information regarding randomisation process. Infants excluded from analysis if 'discharged and failed to complete treatment.'
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	No information on randomisation process.
Allocation concealment (selection bias)	High risk	No information on allocation concealment.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Unblinded study. Prespecified treatment failure criteria, but some subjectivity (e.g. duration of FiO2 increase).
Blinding of outcome assessment (detection bias) Nasal trauma	High risk	Assessment method not stated.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data.
Selective reporting (reporting bias)	High risk	Unclear whether trial registered. Infants excluded from analysis if discharged.
Other bias	Low risk	No other bias noted.

Yoder 2013.

Study characteristics
Methods	Multicentre randomised controlled trial
Participants	432 preterm infants > 28 weeks' GA and BW > 1000 g being managed with non‐invasive respiratory support either as primary support after birth, or postextubation in 7 centres in the USA and China.
Interventions	HFNC (58 preterm infants): various devices starting at 3–5 L/min, increased as required to maximum of 3 L/min above starting point Nasal CPAP (67 preterm infants): 5–6 cmH2O or equivalent to end expiratory pressure on ventilator, subsequently increased to maximum 8 cmH2O
Outcomes	Need for intubation within 72 hours of treatment assignment Duration of respiratory support Delayed intubation Significant apnoea Pulmonary air leaks Feed intolerance Abdominal distension NEC Intestinal perforation Late‐onset nosocomial infection Nasal mucosal injury Infant comfort BPD (based on need for oxygen as assessed by an oxygen reduction test at 36 weeks' PMA) Death Oxygen requirement at discharge Time to full feeds
Notes	Recruitment from December 2007 to April 2012 at multiple centres in the USA and China. Study underpowered because lower incidence than expected of intubation in infants treated with CPAP. No cross‐over permitted between interventions in the first 72 hours of the study. More infants in the HFNC group crossed over to the alternative treatment after 72 hours. 6 Vapotherm devices were provided for use at 3 study sites. No conflicts of interest declared.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random number generation.
Allocation concealment (selection bias)	Low risk	Opaque sealed envelopes.
Blinding of participants and personnel (performance bias)	Unclear risk	Unblinded study.
Blinding of outcome assessment (detection bias) Mechanical ventilation within 72 hours of trial entry	Unclear risk	Unblinded study but prespecified criteria for intubation.
Blinding of outcome assessment (detection bias) Nasal trauma	High risk	Subjective assessment, non‐blinded, criteria not stated.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data.
Selective reporting (reporting bias)	Unclear risk	Some outcomes not reported in detail. Feeding intolerance not reported.
Other bias	Low risk	No other bias noted.

BiPAP: bilevel positive airway pressure; BPD: bronchopulmonary dysplasia; BW: birth weight; CLD: chronic lung disease; CPAP: continuous positive airway pressure; CXR: chest x‐ray; FiO₂: fraction of inspired oxygen; g: gram; GA: gestational age; HF: high flow; HFNC: high flow nasal cannula; INSURE: INtubation, SURfactant, Extubation; IVH: intraventricular haemorrhage; NEC: necrotising enterocolitis; nHF: nasal high flow; NICU: neonatal intensive care unit; NIH: National Institutes of Health; NIMV: nasal intermittent mandatory ventilation; NIPPV: nasal intermittent positive pressure ventilation; PaCO₂: partial pressure of carbon dioxide in arterial blood; pCO₂: partial pressure of carbon dioxide; PDA: patent ductus arteriosus; PEEP: positive end expiratory pressure; PIP: peak inspiratory pressure; PMA: postmenstrual age; PVL: periventricular leukomalacia; RDS: respiratory distress syndrome; ROP: retinopathy of prematurity; SiPAP: synchronised inspiratory positive airway pressure; SpO₂: oxygen saturation.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Akbarian‐Rad 2020	This RCT compared nHF with CPAP following surfactant administration via INSURE (not as primary respiratory support). This study did not examine the use of nHF for the target indication for this review.
Boumecid 2007	This cross‐over trial compared variable flow CPAP with constant‐flow CPAP and non‐humidified nasal cannula at 2 L/min. No outcomes of relevance to this review were recorded.
Campbell 2006	This was a single‐centre study that randomised 40 intubated preterm infants to humidified, unheated nHF (mean gas flow 1.6 L/min) or variable flow CPAP (5–6 cmH2O) after extubation. This study did not examine the use of nHF for the target indication for this review.
Capasso 2005	This study did not examine the use of nasal cannula for the target indication for this review; resuscitation at birth was studied.
Charki 2020	This was a non‐randomised study of nHF versus nasal CPAP in preterm infants following extubation.
Chen 2015	This was an RCT of 66 very low birth weight infants assigned to nHF or nasal CPAP following surfactant therapy. This study did not examine the use of nHF for the target indication for this review.
Chen 2020	This study randomised 94 ELBW infants to nHF or CPAP following extubation within the first 7 days of life. This study did not examine the use of nHF for the target indication for this review.
Collins 2013b	This RCT allocated 132 intubated preterm infants < 32 weeks to nHF or CPAP upon extubation. This study did not examine the use of nHF for the target indication for this review.
Courtney 2001	This cross‐over trial compared variable flow CPAP with constant flow CPAP, and a modified nasal cannula attached to a constant flow CPAP circuit. No outcomes of relevance to this review were recorded.
de Jongh 2014	This study was non‐randomised. It compared work of breathing on CPAP compared with nHF. Initial modality was dependent on what infant was already receiving. No outcomes of relevance to this review were recorded.
Elkhwad 2014	This RCT presented in abstract form allocated ELBW infants to nHF or CPAP following extubation. This study did not examine the use of nHF for the target indication for this review.
Hua 2013	This RCT presented in abstract form randomised infants to nHF or CPAP at day 5 of life. The study did not examine the use of HF for the target indication for this review.
Iranpour 2011	This RCT, published in Persian, that enrolled 70 preterm infants at 24 hours of age following surfactant administration. This study did not examine the use of nHF for the target indication for this review.
Kadivar 2016	This RCT allocated 54 patients to nHF or CPAP following INSURE. This study did not examine the use of nHF for the target indication for this review.
Kang 2016	This RCT allocated infants to nHF or CPAP following 'ventilator weaning'. This study did not examine the use of nHF for the target indication for this review.
Klingenberg 2014	In this RCT, patient comfort was compared between nHF and CPAP. No outcomes of relevance to this review were recorded.
Lampland 2009	This non‐randomised cross‐over study compared CPAP with nHF. No outcomes of relevance to this review were recorded.
Lee 2011	This study presented in abstract form examined nHF versus nasal CPAP following extubation in preterm infants. This study did not examine the use of nHF for the target indication for this review.
Liu 2014	This RCT allocated 255 intubated newborn infants < 7 days of life to nHF or CPAP upon extubation. This study did not examine the use of nHF for the target indication for this review.
Ma 2014	This study examined the use of nHF following extubation. This study did not examine the use of nHF for the target indication for this review.
Manley 2013	This was a non‐inferiority RCT that enrolled very preterm infants to nHF or CPAP after extubation. This study did not examine the use of nHF for the target indication for this review.
Miller 2010	This was a pilot RCT that enrolled preterm infants to 1 of 2 brands of nHF following extubation. This study did not examine the use of nHF for the target indication for this review.
Mostafa‐Gharehbaghi 2015	This RCT allocated infants to nHF or CPAP following surfactant provision via INSURE. This study did not examine the use of nHF for the target indication for this review.
Nasef 2015	Preterm infants < 1500 g were randomised in a cross‐over design to receive 2 hours of either infant flow CPAP (IF‐CPAP) at 5–6 cmH2O or nHF with the flow rate adjusted to achieve an equivalent pharyngeal pressure. No outcomes of relevance to this review were recorded.
Pyon 2008	This cross‐over trial compared nasal CPAP with nHF. No outcomes of relevance to this review were recorded.
Saslow 2006	This cross‐over trial compared CPAP with nHF. No outcomes of relevance to this review were recorded.
Shokouhi 2019	This RCT allocated infants to nHF or CPAP following surfactant provision via INSURE. This study did not examine the use of nHF for the target indication for this review.
Soonsawad 2016	This RCT allocated infants to nHF, or continuing CPAP, for weaning off CPAP. This study did not examine the use of nHF for the target indication for this review.
Soonsawad 2017	This RCT allocated infants to nHF or CPAP following extubation. This study did not examine the use of nHF for the target indication for this review.
Sreenan 2001	This cross‐over trial of CPAP and non‐humidified nHF was non‐randomised.
Wilson 1996	This study examined nasal cannula compared with nasopharyngeal catheters at flow rates < 1 L/min.
Woodhead 2006	This RCT allocated infants to nHF or CPAP following extubation. This study did not examine the use of nHF for the target indication for this review
Zivanovic 2019	This was a retrospective observational study (non‐randomised).

CPAP: continuous positive airway pressure; ELBW: extremely low birth weight; HF: high flow; IF‐CPAP: infant flow CPAP; INSURE: INtubation, SURfactant, Extubation; nHF: nasal high flow; RCT: randomised controlled trial.

Characteristics of studies awaiting classification [ordered by study ID]

Awad 2021.

Methods	RCT
Participants	Preterm infants < 35 weeks' gestation
Interventions	nHF CPAP
Outcomes	Primary outcome Treatment failure Secondary outcomes Mortality Sepsis NEC IVH BPD
Notes

Balasubramanian 2022.

Methods	Single‐centre, double‐blind, RCT
Participants	Preterm infants ≥ 28 weeks' gestation
Interventions	Increased nasal flow therapy (8–10 L/min) Standard nasal flow therapy (5–7 L/min)
Outcomes	Need for higher respiratory support (CPAP or mechanical ventilation) or surfactant therapy
Notes

Cetinkaya 2018.

Methods	RCT
Participants	Preterm infants
Interventions	nHF CPAP NIPPV
Outcomes	Primary outcome Intubation Secondary outcomes Duration of respiratory support Duration of oxygen BPD Air leak ROP IVH
Notes	Attempted to contact authors for further information but received no response.

Febre 2015.

Methods	RCT
Participants	20 preterm and term infants BW 400–5000 g, needing FiO2 > 30%
Interventions	HFNC (quote) "Adaptive Dynamic Inspiratory Nasal Apparatus;" 2–4 L/min, pop‐off valve if circuit pressure exceeds 10 cmH2O CPAP Hudson prongs 4–8 cmH2O
Outcomes	Oxygen requirement Level of pressure or flow support Radiological changes Blood gas measurement Time to wean off protocol, and failure to wean or necessity for endotracheal intubation
Notes	Potential for allocation bias

Iskandar 2019.

Methods	RCT
Participants	Preterm infants 28–35 weeks' gestation
Interventions	nHF CPAP
Outcomes	Treatment failure Length of device use Length of Kangaroo Mother Care Full enteral feeding time Pain score Nasal trauma Systemic complications
Notes

Lawrence 2012.

Methods	RCT
Participants	Infants 26 and 33 6/7 weeks' GA and BW 750–2500 g
Interventions	Nasal CPAP HFNC
Outcomes	Transpleural pressure PDA NEC IVH Air leaks
Notes

Oktem 2021.

Methods	RCT
Participants	Preterm infants < 32 weeks' gestation
Interventions	nHF CPAP NIPPV Nasal high‐frequency oscillatory ventilation
Outcomes	Primary outcome Intubation requirement Secondary outcomes Duration of non‐invasive ventilation Air leak syndrome Abdominal distension IVH NEC Nasal injury Increased secretions Agitation Mortality
Notes

Park 2011.

Methods	No information available
Participants	No information available
Interventions	No information available
Outcomes	No information available
Notes	Abstract submitted to Pediatric Academic Societies conference: no further information available.

Shirvani 2019.

Methods	RCT
Participants	60 infants with RDS < 34 weeks' GA and BW < 2000 g
Interventions	nHF CPAP
Outcomes	Duration of hospitalisation Day to full feeds
Notes	No raw data provided for primary outcomes of interest of review (odds ratios only). No further information available from study authors.

BPD: bronchopulmonary dysplasia; BW: birth weight; CPAP: continuous positive airway pressure; FiO₂; fraction of inspired oxygen; GA: gestational age; HF: high flow; HFNC: high flow nasal cannula; IVH: intraventricular haemorrhage; NEC: necrotising enterocolitis; nHF: nasal high flow; NIPPV: nasal intermittent positive pressure ventilation; PDA: patent ductus arteriosus; RCT: randomised controlled trial; RDS: respiratory distress syndrome; ROP: retinopathy of prematurity.

Characteristics of ongoing studies [ordered by study ID]

ACTRN12610000677000.

Study name	High flow support versus continuous positive airway pressure (CPAP) support in non‐acute respiratory support for preterm infants from 30 weeks' corrected gestation
Methods	RCT 30 infants
Participants	Infants aged ≥ 5 days, ≥ 30 weeks' corrected GA, < 32 weeks' corrected GA, on CPAP in < 5 cmH2O and < 25% oxygen
Interventions	HFNC (4–6 L/min)
Outcomes	CLD Treatment failure Stability of treatment
Starting date	31 August 2010
Contact information	Ashley McEwan (ashley.mcewan@hotmail.com)
Notes	ACTRN12610000677000

ACTRN12611000233921.

Study name	High‐flow for infants in non‐tertiary centres (HINT trial)
Methods	RCT
Participants	Infants > 32 weeks' gestation aged 1–24 hours
Interventions	HF flow rate 6–7 L/min Ambient (head box oxygen)
Outcomes	Treatment failure Transfer to tertiary centre
Starting date	3 March 2011
Contact information	A/Prof Adam Buckmaster (abuckmaster@nsccahs.health.nsw.gov.au)
Notes

CTRI/2017/09/009910.

Study name	High flow nasal cannulae versus nasal continuous positive airway pressure in neonates with respiratory distress syndrome
Methods	Non‐randomised prospective cohort study
Participants	Infants < 36 weeks' GA
Interventions	nHF (Fisher & Paykel Optiflow Junior). Flow rates 1–8 L/min Nasal CPAP (Drager Babylog 8000). CPAP pressure 4–6 cmH2O
Outcomes	Mechanical ventilation within 72 hours of initiation of therapy Days of non‐invasive support Days of oxygen supplementation Nasal injury Pneumothorax Abdominal distension
Starting date	9 February 2015
Contact information	Dr Leslie Lewis (leslielewis1@gmail.com)
Notes	Not prospectively registered

CTRI/2019/10/021633.

Study name	Breathing stabilization in small babies at the time of birth
Methods	RCT 124 infants
Participants	Infants 28–36+6 weeks' GA and BW ≥ 800 g with respiratory distress and FiO2 > 0.3
Interventions	Infants randomised to 1 of 2 groups for delivery room stabilisation nHF (starting flow 4–6 L/min, maximum 8 L/min) Nasal CPAP (starting pressure 5 cmH2O, maximum pressure 8 cmH2O)
Outcomes	Primary outcome Treatment failure Secondary outcomes Time to treatment failure Need for surfactant Duration of respiratory support and supplemental oxygen Age of starting feeds and time to achieve full feeds Duration of hospital stay Weight at discharge Adverse events
Starting date	23 October 2019
Contact information	Sripana Basu (sriparna.neonat@aiimsrishikesh.edu.in)
Notes

Irct2016052510026N.

Study name	Therapeutic effect of heated, humidified, high‐flow nasal cannula (HHHFNC) in respiratory distress syndrome
Methods	Trial design not stated
Participants	Premature infants < 1500 g with RDS
Interventions	NIMV: PIP 16–20 cmH2O, PEEP 5–6 cmH2O, rate 40–50 breaths/min, inspiratory time 0.4 seconds and flow rate 8–10 L/min High flow: 2.5–3 L/min Surfactant if FiO2 > 0.3 via INSURE
Outcomes	Duration of non‐invasive respiratory support
Starting date	Not stated
Contact information	Not stated
Notes

Irct20180226038865N.

Study name	Comparison of two forms of non‐invasive respiratory supporting preterm infants
Methods	Trial design not stated
Participants	Preterm infants with RDS requiring surfactant
Interventions	Nasal CPAP Humidified HFNC
Outcomes	Response to treatment
Starting date	Not stated
Contact information	Not stated
Notes

Irct20190623043988N.

Study name	Continuous positive airway pressure or high flow nasal cannula for respiratory distress syndrome
Methods	RCT
Participants	Preterm infants BW < 2000 g and < 34 weeks' gestation
Interventions	nHF 2–5 L/min CPAP 4–6 cmH2O
Outcomes	Intervention 'effectiveness'
Starting date
Contact information
Notes

Irct20200616047788N.

Study name	Comparing two respiratory support methods in RDS treatment of premature neonates
Methods	RCT
Participants	Preterm infants 28–34 weeks' gestation
Interventions	nHF CPAP
Outcomes	Primary outcome RDS score Secondary outcomes Air leak syndromes BPD IVH Intubation PDA
Starting date
Contact information
Notes

ISRCTN66716753.

Study name	Can high flow nasal prongs therapy facilitate earlier establishment of full oral feeds in babies who are nasal continuous positive airway pressure dependent at 32 weeks gestation?
Methods	RCT 44 infants
Participants	Infants aged < 30 weeks and BW < 1500 g requiring CPAP at 32 weeks' corrected age with oxygen requirement < 30%
Interventions	Continue on nasal CPAP HF nasal prongs 7 L/min
Outcomes	Establishment of full oral feeding
Starting date	February 2013
Contact information	Dr Jan Miletin (jmiletin@coombe.ie)
Notes	ISRCTN66716753

NCT01270581.

Study name	High flow nasal cannula vs bubble nasal CPAP for the treatment of transient tachypnoea of the newborn in infants > 35 weeks gestation
Methods	RCT Estimated enrolment 66 infants
Participants	Infants > 35 weeks' gestation diagnosed with transient tachypnoea and admitted to NICU within first 24 hours of life
Interventions	HFNC Bubble nasal CPAP
Outcomes	Duration of respiratory support
Starting date	July 2010
Contact information	Andrea Weintraub, Mount Sinai School of Medicine, New York (andrea.weintraub@mssm.edu)
Notes	ClinicalTrials.gov identifier: NCT01270581

NCT02055339.

Study name	Comparison of nasal continuous positive airway pressure with low flow oxygen versus heated, humidified high flow nasal cannula for oral feeding of the premature infant (CHOMP Trial): a pilot study
Methods	RCT Sample size 40 infants
Participants	Preterm infants born < 28 weeks' gestation who were then 34 weeks' corrected GA, dependent on non‐invasive respiratory support, and receiving nasogastric feeds.
Interventions	HFNC (Fisher & Paykel) Infant flow CPAP
Outcomes	Time to reach full oral feeds
Starting date	February 2014
Contact information	Sandra Leibel, Mount Sinai Hospital, New York (sleibel@mtsinai.on.ca)
Notes	ClinicalTrials.gov Identifier: NCT02055339

NCT02499744.

Study name	Humidified high flow nasal cannula versus nasal intermittent positive ventilation in neonates as primary respiratory support: a randomised controlled trial
Methods	Randomised trial
Participants	Preterm infants > 28 weeks' GA and BW > 1000 g with respiratory distress
Interventions	High flow (2–8 L/min) NIPPV (12–22 cmH2O/5–7 cmH2O)
Outcomes	Endotracheal intubation within 72 hours
Starting date	February 2016
Contact information	Gao WeiWei
Notes

UMIN000018983.

Study name	Effective and safe use of heated humified high flow nasal cannula in neonates
Methods	RCT
Participants	Infants > 28 weeks' GA and BW > 1000 g with neonatal 'respiratory disorder' not responsive to initial nasal CPAP therapy
Interventions	Nasal CPAP (no details given) nHF (no details given)
Outcomes	Improvement in respiratory acidosis and hypercapnia 1 hour after intervention Improvement in oxygenation 1 hour after intervention
Starting date	1 November 2015
Contact information	Atsuko Taki (ataki.ped@tmd.ac.jp)
Notes	Planned recruitment of 20 infants

BPD: bronchopulmonary dysplasia; BW: birth weight; CLD: chronic lung disease; CPAP: continuous positive airway pressure; FiO₂: fraction of inspired oxygen; g: gram; GA: gestational age; HF: high flow; HFNC: high‐flow nasal cannula; INSURE: INtubation, SURfactant, Extubation; IVH: intraventricular haemorrhage; nHF: nasal high flow; NICU: neonatal intensive care unit; nasal intermittent mandatory ventilation; NIPPV: nasal intermittent positive pressure ventilation; PDA: patent ductus arteriosus; PEEP: positive end expiratory pressure; PIP: peak inspiratory pressure; RCT: randomised controlled trial; RDS: respiratory distress syndrome.

Differences between protocol and review

• For the 2022 review update, we added the methodology and plan for summary of findings tables and GRADE recommendations, which were not included in the previous versions of the review. These changes were to comply with the latest Cochrane standards.

• For the 2022 review update, we developed a new search strategy, which we ran without date limits (Appendix 1).

• The protocol and earlier versions of this review included Embase as a search source. This update omitted Embase based on the rationale that CENTRAL now includes records from Embase. This rationale has, subsequently, been flagged as a method which may reduce sensitivity of the search. Subsequent updates will include Embase as a source.

• The previous versions of this review included the use of nHF for other indications in preterm infants (Wilkinson 2011; Wilkinson 2016). Due to the increase in number of studies of nHF in preterm infants, this review is limited to studies of nHF as primary respiratory support. The title has been updated to reflect this (previous title was 'High flow nasal cannula for respiratory support in preterm infants'). A second review (yet to be completed) will examine nHF for postextubation or postsurfactant respiratory support.

• We updated the Criteria for considering studies for this review section to clarify that studies with a superiority, non‐inferiority or equivalence hypothesis could be included.

• The term 'chronic lung disease' has been changed to 'bronchopulmonary dysplasia'.

• The previous versions of this review also compared nHF with high flow delivered with alternative devices; this has been removed from the current version of the review.

• Some changes have been made to the outcome measures from Wilkinson 2016 to this update:

  • treatment failure and mechanical ventilation have been made primary outcomes (were secondary outcomes in the previous review);

  • treatment failure and mechanical ventilation have been defined as occurring within 72 hours of trial entry (seven days in previous review);

  • surfactant administration (via any method) was added as a secondary outcome.

• The following headings or text were added or updated:

  • unit of analysis;

  • dealing with missing data;

  • assessment of heterogeneity;

  • assessment of reporting biases;

  • data synthesis;

  • potential biases in the review process;

  • agreements and disagreements with other studies or reviews;

  • summary of findings and assessment of certainty of the evidence.

• The summary of findings tables were added.

• Subgroups analyses were not present in the protocol, but were added to Wilkinson 2011 and modified in Wilkinson 2016.

• The subgroup analyses (for the primary outcomes) added with this update were:

  • surfactant administration permitted during the intervention period (without this being deemed treatment failure) versus not permitted;

  • second‐line CPAP permitted in the nHF arm prior to mechanical ventilation versus not permitted.

Contributions of authors

The protocol was developed by Wilkinson, Andersen and O'Donnell (Wilkinson 2007).

Kate Hodgson and Brett Manley conducted screening for the 2022 update, and performed the GRADE assessment, with any uncertainty resolved by Dominic Wilkinson.

For the articles in which BJM had a conflict of interest due to authorship, KH and DW performed the GRADE assessment.

All authors (KH, BJM, AGDP, DW) provided data analysis and collaborated in the writing of the review.

For the articles in which BJM had a conflict of interest due to authorship, the other authors (KH, AGDP, DW) performed the data analysis.

Sources of support

Internal sources

• University of Oxford, UK

  DW has University of Oxford affiliation.

• Royal Women's Hospital, Australia

  KH and BM are employed by the Royal Women's Hospital, Melbourne.

• University of Melbourne, Australia

  BM is an Associate Professor at the University of Melbourne. KH is undertaking a PhD through the University of Melbourne.

• NHMRC, Australia

  KH has received salary support via a National Health and Medical Research Council (NHMRC) Australia grant. Her salary support includes funding for A/Prof Brett Manley (supervisor) via NHMRC.

• University of Melbourne, Australia

  KH received support from the University of Melbourne Henry and Rachel Ackman Travel Scholarship for attendance at the jENS Congress 2019.

• NHMRC, Australia

  BJM received a Government career development grant fellowship from the Medical Research Future Fund (Australia), and a project grant from NHMRC (Australia).

External sources

• NHMRC, Australia

  Fellowship ‐ DW, BM

• 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

KH works as a Neonatologist, Royal Women's Hospital, Royal Children's Hospital (PIPER retrieval). She is PhD candidate, currently undertaking a PhD in neonatal high flow. Her PhD main area of study is an RCT of high flow during neonatal endotracheal intubation; equipment for this is supplied by Vapotherm. Vapotherm had no input into the trial design, nor access to trial data or the manuscript prior to publication. This trial is not eligible for inclusion in this review. KH has been a co‐author on review articles and book chapters which include descriptions of nasal high flow.

BJM has published several original research articles, review articles and editorials on the topic of nasal high flow in peer‐reviewed journals. He works as a Consultant Neonatologist at the Royal Women's Hospital, Parkville, Victoria, Australia. BJM is one of KH's PhD supervisors, and is a co‐investigator on the RCT of high flow during neonatal endotracheal intubation. This trial is not eligible for inclusion in this review. BM was an author of two of the trials included in this review (Manley 2019; Roberts 2016). These studies were funded by NHMRC (Australia). Analysis of those papers was performed by other review authors (DW, AGDP and KH).

AGDP works as a Consultant Neonatologist, Royal Hobart Hospital, Tasmania, Australia.

DW works as a Consultant Neonatologist, John Radcliffe Hospital, Oxford, UK.

New search for studies and content updated (conclusions changed)