Continuous positive airway pressure (CPAP) for apnoea of prematurity

Abstract

Background

Apnoea of prematurity (AoP) is defined as a pause in breathing for 20 seconds or longer, or for less than 20 seconds when accompanied by bradycardia and hypoxaemia, in a preterm infant. An association between the severity of apnoea and neurodevelopmental delay has been reported. Continuous positive airway pressure (CPAP) is a form of non‐invasive ventilatory assistance that has been shown to be relatively safe and effective in preventing and treating respiratory distress among preterm infants. It is less clear whether CPAP treatment is safe and effective in the prevention and treatment of AoP.

Objectives

1. To assess the effects of CPAP on AoP in preterm infants (this may be compared to supportive care or mechanical ventilation).
2. To assess the effects of different CPAP delivery systems on AoP in preterm infants.

Search methods

Searches were conducted in September 2022 in the following databases: Cochrane Library, MEDLINE, Embase, and CINAHL. We also searched clinical trial registries and the reference lists of studies selected for inclusion.

Selection criteria

We included all randomised and quasi‐randomised controlled trials (RCTs) in which researchers determined that CPAP was necessary for AoP in preterm infants (born before 37 weeks). Cross‐over studies were also included, provided sufficient data were available for analysis.

Data collection and analysis

We used the standard methods of Cochrane and Cochrane Neonatal, including independent assessment of risk of bias and extraction of data by at least two review authors. Discrepancies were resolved by involvement of a third author. We used the GRADE approach to assess the certainty of evidence for the following outcomes: 1) failed CPAP; 2) apnoea; 3) adverse effects of CPAP.

Main results

We included four single‐centre trials conducted in Malaysia, Spain, Germany, and North America, involving 138 infants with a mean/median gestation of 26 to 28 weeks. Two studies were parallel‐group RCTs and two were cross‐over trials. None of the studies compared CPAP with supportive care. All trials compared one form of CPAP with another. Two compared a variable flow device with ventilator CPAP, one compared two different variable flow devices, and one compared a variable flow device with bubble CPAP. Interventions were administered for periods ranging between six and 48 hours, with pressures between 4 and 6 cm H₂O. We assessed all trials as having a high risk of bias for blinding of participants and personnel, and two studies for blinding of outcome assessors. We found a high risk of a carry‐over effect in two studies where the washout period was not adequately described, and a high risk of bias in a study that appeared to use an analysis method not generally accepted for cross‐over studies.

Comparison 1. CPAP and supportive care compared to supportive care alone

We did not identify any study for inclusion in this comparison.

Comparison 2. CPAP delivered by different types of devices

2a. Variable flow compared to ventilator CPAP

Two studies were included in this comparison. We are very uncertain whether there is any difference in the incidence of failed CPAP, defined as the need for mechanical ventilation (risk ratio (RR) 0.16, 95% confidence interval (CI) 0.01 to 2.90; 1 study, 26 participants; very low‐certainty). We are very uncertain whether there is any difference in the frequency of apnoea events (mean difference (MD) per four‐hour interval ‐0.10, 95% CI ‐1.30 to 1.10; 1 study, 26 participants; very low‐certainty). We are uncertain whether there is any difference in adverse events. Neurodevelopmental outcomes were not reported.

2b. Variable flow compared to bubble CPAP

We included one study in this comparison, but it did not report our pre‐specified outcomes.

2c. Infant Flow variable flow CPAP compared to Medijet variable flow CPAP

We are very uncertain whether there is any difference in the incidence of failed CPAP (RR 2.62, 95% CI 0.91 to 7.53; 1 study, 80 participants; very low‐certainty). The frequency of apnoea was not reported, and we do not know whether there is any difference in adverse events. Neurodevelopmental outcomes were not reported.

Comparison 3. CPAP compared to mechanical ventilation

We did not identify any studies for inclusion in this comparison.

Authors' conclusions

Due to the limited available evidence, we are very uncertain whether any CPAP device is more effective than other forms of supportive care, other CPAP devices, or mechanical ventilation for the prevention and treatment of AoP. The devices used in these studies included two types of variable flow CPAP device: bubble CPAP and ventilator CPAP. For each comparison, data were only available from a single study. There are theoretical reasons why these devices might have different effects on AoP, therefore further trials are indicated.

Plain language summary

Continuous positive airway pressure for apnoea of prematurity

Key messages

We did not find any study that compared CPAP with supportive care for apnoea of prematurity.

We found four studies comparing one type of CPAP to another. The studies were small and the care given by healthcare workers may have been influenced by knowing which treatment the baby was receiving. Therefore, we don’t know from these studies whether there is any difference between different types of CPAP for reducing apnoea in babies born too early.

What is apnoea of prematurity?

Babies born early (preterm) may experience prolonged pauses in their breathing. In a baby born too early, we call such pauses apnoea of prematurity. Apnoea may be due to immaturity of the mechanisms controlling breathing or to weakness of the muscles that keep their airway open during breathing. In severe apnoea the baby’s heart rate will slow and there will be a drop in the level of oxygen. If this happens too often it may affect the baby’s brain development. When babies have apnoea, supportive care is usually given. This may be in the form of stimulation, by stroking the baby or moving their limbs, providing oxygen, or both.

What is CPAP?

Continuous positive airway pressure (CPAP) is a form of breathing support, often referred to as a type of non‐invasive ventilation. The continuous pressure is applied usually through a nose piece, allowing the baby to continue to breathe by themselves but making breathing easier. This is probably because the positive pressure that is applied helps to keep the airway open.

What did we want to find out?

We wanted to find out whether using CPAP in preterm babies might reduce apnoea or the need for a more invasive form of treatment such as mechanical ventilation, i.e. using a machine to do the breathing for the baby. We also wanted to know whether any particular method of applying CPAP was more effective.

What did we do?

We searched for studies that included preterm babies where the researchers thought CPAP was necessary for apnoea of prematurity. We included studies that compared CPAP with supportive care (such as stimulation, oxygen, or both) or studies that compared different types of CPAP.

What did we find?

We did not find any studies that compared CPAP with supportive care. We found four small studies, involving 138 babies. They compared two different forms of CPAP. There were three comparisons with only one study in each comparison and one study did not provide any useful data. The duration of CPAP in the four studies was between 4 and 48 hours. From these four studies, we do not know whether there is any difference between different forms of CPAP. This is because each comparison had only one study, the results were very imprecise because the studies were small, and the care given by the healthcare workers might have been influenced by their awareness of which treatment the baby was receiving.

How up‐to‐date is the evidence?

The evidence is up‐to‐date to 6 September 2022.

Summary of findings

Summary of findings 1. CPAP and supportive care compared to supportive care for apnoea of prematurity.

Patient or population: preterm infants with apnoea of prematurity Setting: neonatal units Intervention: CPAP and supportive care (such as oxygen therapy) Comparison: supportive care alone
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with supportive care	Risk with CPAP
Failed CPAP	No study reported this outcome				—	—
Mean apnoea events	No study reported this outcome				—	—
Adverse events	No study reported this outcome				—	—
Long‐term neurodevelopmental outcomes	No study reported this outcome				—	—
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CPAP: continuous positive airway pressure
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

Summary of findings 2. Variable flow compared to ventilator CPAP for apnoea of prematurity.

Patient or population: preterm infants with apnoea of prematurity (median age 9 days) Setting: single neonatal intensive care unit Intervention: variable flow CPAP Comparison: ventilator CPAP
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with conventional ventilator CPAP	Risk with variable flow
Failed CPAP (defined as need for mechanical ventilation)	Study population		RR 0.16 (0.01 to 2.90)	26 (1 RCT)	⊕⊝⊝⊝ VERY LOW1,2,3	—
	214 per 1000	34 per 1000 (2 to 621)
Mean apnoea events (apnoea > 20 seconds with or without bradycardia or desaturation) Mean per 4‐hour interval	The mean apnoea events was 2.6 events in a 4‐hour interval	MD 0.1 lower (1.3 lower to 1.1 higher)	—	26 (1 RCT)	⊕⊝⊝⊝ VERY LOW1,2,3	—
Long‐term neurodevelopmental outcomes	No study reported this outcome				—	Other adverse events were not reported
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CPAP: continuous positive airway pressure; MD: mean difference; 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

¹Downgraded for lack of blinding.

²Downgraded by one level because data from only one small study in a single setting.

³Downgraded by two levels because confidence interval includes important benefit and important harm.

Summary of findings 3. Variable flow compared to bubble CPAP for apnoea of prematurity.

Patient or population: preterm infants, 24 to 29 weeks gestation and mean age 18 days with apnoea of prematurity Setting: neonatal unit Intervention: variable flow Comparison: bubble CPAP
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with bubble CPAP	Risk with variable flow
Failed CPAP	No study reported this outcome				—	—
Mean apnoea events	No study reported this outcome				—	This outcome was reported by 1 study as number of infants with resolution of apnoea at 24 hours, and reported no difference between the groups
Adverse events	No adverse events reported during study period				—	—
Long‐term neurodevelopmental outcomes	No study reported this outcome				—	—
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CPAP: continuous positive airway pressure
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

Summary of findings 4. Infant Flow CPAP compared to Medijet CPAP for apnoea of prematurity.

Patient or population: preterm infants with apnoea of prematurity (gestation 27 weeks, age not stated) Setting: neonatal unit Intervention: Infant Flow variable flow CPAP Comparison: Medijet variable flow CPAP
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with Medijet variable flow CPAP	Risk with Infant Flow variable flow CPAP
Failed CPAP	Study population		RR 2.62 (0.91 to 7.53)	80 (1 RCT)	⊕⊝⊝⊝ VERY LOW1,2,3	—
	103 per 1000	269 per 1000 (93 to 772)
Mean apnoea events	This outcome was not reported. Authors reported no difference in the number of infants with resolution of apnoea at 24 hours.					—
Adverse events	Severe abdominal distension reported for 2 participants in each group					Pneumothorax was reported, but data not presented separately by study group
Long‐term neurodevelopmental outcomes	No study reported this outcome					—
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CPAP: continuous positive airway pressure; 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

¹Downgraded by two levels for very serious concerns about lack of blinding of interventions and cross‐over design with concerns about a carry‐over effect.

²Downgraded by one level because data from only one study in a single setting.

³Downgraded by one level because confidence interval incorporates both benefit and harm and small information size.

Background

Description of the condition

Apnoea of prematurity (AoP) is widely defined as a pause in breathing for 20 seconds or longer, or for less than 20 seconds when accompanied by bradycardia and hypoxaemia, in a preterm infant (Eichenwald 2016). Incidences of apnoea are inversely related to gestational age, affecting almost all infants of 28 weeks’ gestation or less and typically resolving by 40 weeks' postmenstrual age (Henderson‐Smart 1981). An association between the severity of apnoea and neurodevelopmental delay has been reported (Janvier 2004; Pillekamp 2007).

Traditionally, AoP has been classified into three groups: central, obstructive, and mixed apnoea. Central apnoea is caused by immaturity, or instability, of the respiratory feedback mechanism within the central nervous system (Eichenwald 2016). Obstructive apnoea is caused by closure of the upper airway, such as collapse or instability of pharyngeal tissue, nasal obstruction and, in some cases, neck flexion; the infant continues to breathe against the obstruction, thus obstructive apnoea is distinct from central apnoea due to the presence of chest motion. Mixed apnoea, the most common form, as its name suggests combines characteristics of both central and obstructive apnoea. Central apnoea can precede or follow obstructive apnoea.

Description of the intervention

Continuous positive airway pressure (CPAP) is a form of non‐invasive ventilatory assistance that is applied as constant pressure throughout the respiratory cycle, usually via the nostrils.

CPAP has an established role in the management of respiratory distress in preterm infants (Ho 2020a; Ho 2020b; Subramaniam 2021), and it is used to facilitate weaning from mechanical ventilation (Davis 2003).

Nasal CPAP can be applied via a binasal prong, a single nasopharyngeal prong, or a nasal mask. In the past, CPAP has been applied via face mask or by endotracheal tube (Davis 2001b; Rhodes 1973). Evidence suggests that the binasal prong is more effective than the single nasopharyngeal prong when used in weaning from mechanical ventilation (De Paoli 2008). Pressure, which is generated by a bubble circuit, a mechanical ventilator, or a CPAP driver, is generally set at a constant level, commonly between 4 and 8 cmH_₂O (Bamat 2021; Eichenwald 2016). However, higher pressures have been used (Krouskop 1975).

Infants at risk of AoP would usually be monitored with an apnoea or cardiorespiratory monitor, which might include oxygen saturation. For apnoeic episodes that are not clinically significant or frequent, simple interventions such as proper positioning and physical stimulation may be sufficient. In addition, oxygen therapy might be initiated. Blood transfusion is usually administered in the presence of anaemia. Methylxanthines might also be given to prevent or treat apnoea (Henderson‐Smart 2010a; Henderson‐Smart 2010b). They may be used prior to initiating CPAP, concurrently with CPAP, or as an additional treatment when CPAP fails (Eichenwald 2016). Methylxanthines (including theophylline, aminophylline, and caffeine) reduce apnoea and the need for mechanical ventilation. In addition, caffeine improves infant growth and long‐term neurodevelopmental outcomes, and is less toxic (Henderson‐Smart 2010b).

CPAP is typically considered after an episode of clinically significant apnoea or after frequent mild apnoeic episodes. The effectiveness of nasal intermittent positive pressure ventilation (NIPPV) compared to nasal CPAP for apnoea of prematurity has been studied. Trial results suggest that NIPPV may be more effective than nasal CPAP for preterm babies with severe or frequent apnoeic episodes (Lemyre 2017). If apnoea were deemed to be persistent, then endotracheal intubation and mechanical ventilation might be instituted.

How the intervention might work

In general, CPAP works by opening up collapsed airways, thus increasing functional residual capacity (FRC). This results in a decrease in ventilation perfusion (V/Q) mismatch, hence improving gas exchange within the lung (Rodriguez 2002).

The mechanism of how CPAP prevents central apnoea might be described as stimulation of pulmonary stretch receptors in the lungs and sending of feedback to the respiratory centre of the brain, thus providing a respiratory drive (Hering‐Breuer reflex) (Speidel 1976). CPAP might also stabilise the chest wall, thus reducing inhibitory feedback from the chest wall reflexes during rapid eye movement (REM) sleep (Hagan 1977), and it may regularise respiration during periodic breathing (Kattwinkel 1975). CPAP might reduce the obstructive component of AoP by splinting the pharyngeal airway, thus improving the patency of the upper airway (Miller 1985).

There are two main types of CPAP systems. The first consists of variable flow devices, which include flow drivers such as a specific nasal prong or mask. The second type, such as bubble CPAP systems and infant ventilators, consists of a continuous flow device (Courtney, 2007). Although all CPAP systems have their own specific properties, it is not clear whether there are any clinically important differences between them (De Paoli 2008). Nevertheless, they could vary in their effectiveness in preventing or treating AoP. For example, the bubbling mechanism in bubble CPAP affects a process described as facilitated diffusion caused by pressure oscillations from the bubbles (Pillow 2007). Animal studies suggest that measures of gas exchange are better when compared with ventilator CPAP (Huang 2008). Variable flow CPAP generates pressure at the airway proximal to the infant’s nares via a Bernoulli effect to maintain a constant pressure. During the expiratory phase in variable CPAP, the fluidic flip mechanism causes the flow to flip around to enable preterm babies to exhale with less effort (Klausner 1996), possibly making apnoea less likely. Variable flow CPAP also maintains a better uniform pressure level compared to continuous flow CPAP.

CPAP may be less effective for premature babies of lower gestation and birth weight because these infants may have less mature respiratory systems and more compliant chest walls (Ammari 2005).

When compared with invasive ventilatory methods, CPAP is generally free of adverse effects, the most common of which is trauma to the nares. This can range from mild oedema to septal necrosis (Fischer 2010). Pneumothorax has also been reported (Ho 2020b). The use of oxygen to treat AoP can result in prolonged hyperoxia, a potential cause of retinopathy of prematurity (van Zanten 2014).

Why it is important to do this review

CPAP has been shown in several reviews to be relatively safe and effective in preventing and treating respiratory distress among preterm infants, and in reducing the adverse effects associated with ventilation (Ho 2020a; Ho 2020b; Subramaniam 2021). In addition, CPAP has been compared with nasal intermittent positive pressure ventilation for the treatment of respiratory distress (Lemyre 2016; after extubation, Lemyre 2017), and for AoP (Lemyre 2017).

However, it is important to determine whether CPAP is safe and effective for the prevention and treatment of AoP, and whether delivery by any particular CPAP device is superior.

Objectives

1. To assess the effects of CPAP on AoP in preterm infants (this may be compared to supportive care or mechanical ventilation).

2. To assess the effects of different CPAP delivery systems on AoP in preterm infants.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised and quasi‐randomised controlled trials looking at treatment of AoP via CPAP. Cross‐over studies reporting short‐term outcomes such as apnoea frequency were also included, provided sufficient data were available for analysis.

Types of participants

We included trials of preterm infants (born before 37 weeks) in which the investigator had determined that CPAP is necessary for AoP. The criteria for determining whether CPAP is required were those defined by trial investigators.

Types of interventions

We included the following comparisons.

1. CPAP and supportive care compared to supportive care alone.

2. CPAP delivered by different types of devices (we compared one CPAP device versus any other, e.g. ventilator CPAP, bubble, or driver CPAP).

3. CPAP compared to mechanical ventilation.

Supportive care may include physical stimulation, the use of oxygen, or both. Blood transfusion may also be given in the presence of anaemia.

For any of these comparisons, medication or another treatment for AoP could be used as a co‐intervention or as back‐up treatment for CPAP, provided it was used in both comparison groups.

Trials comparing bilevel positive airway pressure (BiPAP) with conventional nasal CPAP were excluded because BiPAP is considered to be a form of nasal intermittent positive‐pressure ventilation (NIPPV), which is covered in another Cochrane Review (Lemyre 2017).

Trials not including apnoea as an outcome were not included. Studies on CPAP used for respiratory distress were not included, even if apnoea was an outcome.

Types of outcome measures

Primary outcomes

1. Failed CPAP: defined as any form of apnoea that is sufficiently frequent or severe as to require additional ventilator support, such as NIPPV or endotracheal intubation, or backup medical therapy such as a methylxanthine (for comparisons 1 and 2). The criteria used would be those defined by study authors (such as more than two apnoeic episodes within a four‐hour period requiring bag and mask ventilation)

2. Apnoea

   1. Rates of apnoea and bradycardic events expressed as events per specified time period (apnoea with or without bradycardia or bradycardia alone)

   2. Differences in rates of apnoea (before and after treatment) expressed as events per hour (for all comparisons)

Secondary outcomes

1. Use of methylxanthines as backup treatment

2. Rate of apnoea with bradycardia expressed as events per specified time period

3. Bronchopulmonary dysplasia (oxygen requirement at 36 weeks' postmenstrual age)

4. Mortality

5. Pneumothorax (confirmed by chest X‐ray)

6. Retinopathy of prematurity (any stage and > stage 2)

7. Rate of gastrointestinal (GI) complications (abdominal distension requiring cessation of feeds, GI perforation diagnosed radiologically or at operation, necrotising enterocolitis)

8. Long‐term neurodevelopmental outcomes (neurodevelopment measured on a validated scale that measures cognitive, motor, or behavioural function, or blindness, deafness, or cerebral palsy, at about two years of age)

Search methods for identification of studies

The search strategies were written by the Cochrane Information Specialist. The search strategy combined intervention terms (CPAP) with apnoea, and neonatal terms standardised by Cochrane Neonatal.

Electronic searches

Searches for trials were conducted without date, language, or publication type limits. Searches for systematic reviews were limited from 2020 forward. The following databases were searched:

• Cochrane Library (CENTRAL and CDSR) (2022, Issue 8), Wiley (6 September 2022);

• Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process, In‐Data‐Review & Other Non‐Indexed Citations, Daily and Versions 1946 to 2 September 2022;

• Embase 1974 to 2 September 2022;

• CINAHL Complete (EbscoHost) (1975 to 06 September 2022).

Search strategies for each database are provided in Appendix 1; Appendix 2; Appendix 3; Appendix 4.

Searching other resources

We searched the following trial registries in June and September 2022:

• US National Library of Medicine (https://www.clinicaltrials.gov);

• ISRCTN Registry (https://www.isrctn.com);

• International Clinical Trials Registry Platform (ICTRP) (https://trialsearch.who.int).

Search strategies are provided in Appendix 5.

We reviewed the reference lists of all the full‐text articles we retrieved and other related systematic reviews. We also approached experts in the field.

We searched PubMed for retractions and errata on 23 September 2022 (PubMed 2022).

Data collection and analysis

We used the standard methods of Cochrane Neonatal.

Selection of studies

We identified and excluded duplicates. If we had found multiple reports of the same study we intended to collate them so that each study rather than each report was the unit of interest in the review. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram (Page 2021), along with Characteristics of included studies and Characteristics of excluded studies tables.

Data extraction and management

Two review authors (EL and JWL) screened the title and abstract of all studies identified by the search strategy, and two review authors (JJL and IZG) independently assessed the full articles for all potentially relevant trials. We excluded studies that did not meet all inclusion criteria, and we have stated the reason for exclusion. We discussed any disagreements until consensus was reached. Two review authors independently extracted the data from the included studies (JJL and IZG). We used a standard data collection form, adapted for the purpose of collecting all specified outcome measures. We resolved discrepancies through discussion with a third review author (JJH).

Assessment of risk of bias in included studies

Two review authors (JJL and JWL) independently assessed the risk of bias (low, high, or unclear) of all included trials using the Cochrane risk of bias tool for the following domains (Higgins 2017).

1. Sequence generation (selection bias).

2. Allocation concealment (selection bias).

3. Blinding of participants and personnel (performance bias).

4. Blinding of outcome assessors (detection bias).

5. Incomplete outcome data (attrition bias).

6. Selective reporting (reporting bias).

7. Any other bias.

We resolved any disagreements by discussion or by consultation with a third review author (JJH). See Appendix 6 for a more detailed description of risk of bias assessment for each of these domains.

Measures of treatment effect

We analysed treatment effects in individual trials using Review Manager 5.4 (Review Manager 2020), and we reported risk ratio (RR) and risk difference (RD) for dichotomous data, as well as mean difference (MD) for continuous data, with respective 95% confidence intervals (CIs). We would have determined the number needed to treat for an additional beneficial outcome (NNTB) or an additional harmful outcome (NNTH) if we had encountered analyses with a statistically significant difference in RD. If studies had reported rates of apnoea per infant per unit of time, we would have reported these as rate ratios.

Unit of analysis issues

The unit of analysis we used was the participating preterm infant in individually randomised trials and cross‐over trials. We considered each infant only once in an analysis. We included cross‐over studies if they measured only short‐term outcomes, such as apnoea frequency, and provided sufficient data for analysis. Significant effects on respiratory rate and oxygenation have been found to continue for up to six hours after discontinuation of CPAP (Abdel‐Hady 1998). Therefore, to avoid a carry‐over effect, we had planned to exclude studies with a washout period of less than six hours. If the washout period was unclear, we included these studies and judged them to be at high risk of bias for a carry‐over effect. We did not find any cluster‐randomised controlled trials (RCTs), but if we did encounter any, we would have followed the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019).

Dealing with missing data

We contacted the authors of two studies with unclear methods (Kattwinkel 1975; Telenko 1999), and in both instances the authors were not able to provide additional information.

When data were missing and could not be derived, we approached analysis of missing data as follows.

1. We would have contacted the original study investigators to request missing data when more than 20% of data were missing (Guyatt 2017).

2. We would have standardised events reported to a commonly used period for mean apnoea rate.

3. If different time periods were used for mean apnoea rate, we would have contacted the study authors and requested data for events per hour, along with standard deviation (SD).

4. If this was not possible, we would have considered converting apnoea rates to events per hour by borrowing an SD from another similar study if appropriate.

Assessment of heterogeneity

If we had been able to perform a meta‐analysis we would have assessed clinical heterogeneity by visually inspecting forest plots and statistical heterogeneity by using the I² statistic. We would have calculated the I² statistic for each RR analysis to quantify inconsistency across studies and to describe the percentage of variability in effect estimates that may be due to heterogeneity rather than to sampling error. We would have defined heterogeneity as follows: I² statistic less than 25%: none; I² statistic 25% to 49%: low; I² statistic: 50% to 74%: moderate; I² statistic 75% or greater: high (Higgins 2019). If we had detected 'high' levels of heterogeneity (I² statistic of 75% or greater), we would explore possible causes (e.g. differences in study design, participants, interventions, or completeness of outcome assessments).

Assessment of reporting biases

We would have assessed publication bias by visually inspecting funnel plot asymmetry if our meta‐analyses had included at least 10 studies. If we had found significant asymmetry in the funnel plot, we would have reported this in the corresponding results.

Data synthesis

We used the statistical package, Review Manager 5 (Review Manager 2020). We would have used fixed‐effect meta‐analysis for combining data when it was reasonable to assume that studies were estimating the same underlying treatment effect (i.e. when trials were examining the same intervention, reported outcomes were sufficiently similar, and trial populations and methods were judged sufficiently similar). If rates were reported, we would have used generic inverse variance meta‐analysis to analyse rates of apnoea.

Subgroup analysis and investigation of heterogeneity

If there were enough studies, we would have undertaken subgroup analyses according to:

1. birth weight group: > 1500 g, 1001 g to 1500 g, and 1000 g or below;

2. gestation group: 34 to 37, 28 to 33, or < 28 weeks' gestation;

3. no apnoea medication versus concomitant medication;

4. different CPAP interfaces (such as nasal mask, binasal prong, single nasal prong) (for comparisons 1 and 2); and

5. different definitions of failed treatment (more severe apnoea versus less severe apnoea or backup treatment ventilation versus medication).

Sensitivity analysis

We planned to undertake sensitivity analyses to determine whether review findings were affected by including only studies using adequate methods (low risk of bias), defined as adequate randomisation and allocation concealment, blinding of intervention and outcome measurement, and less than 10% loss to follow‐up.

Summary of findings and assessment of the certainty of the evidence

We used the GRADE approach, as outlined in the GRADE Handbook (Schunemann 2013), to assess the certainty of evidence for the following (clinically relevant) outcomes for each comparison.

1. Failed CPAP: defined as apnoea that is sufficiently frequent or severe as to require additional ventilator support, including endotracheal intubation.

2. Apnoea.

   1. Rates of apnoea and bradycardia expressed as events per hour.

   2. Differences in rates of apnoea (before and after treatment) expressed as events per hour.

3. Adverse effects of CPAP (pneumothorax, gastrointestinal effects).

4. Long‐term neurodevelopmental outcomes (neurodevelopment measured on a validated scale that measures cognitive, motor, or behavioural function, or blindness, deafness, or cerebral palsy, at about two years of age).

Two review authors (JJH and IZG) independently assessed the certainty of evidence for each of the outcomes above. We considered evidence from RCTs as high certainty and downgraded the evidence by one level for serious (or two levels for very serious) limitations based upon the following: design (risk of bias), consistency across studies, directness of evidence, precision of estimates, and presence of publication bias. We used GRADEpro GDT to create Table 1, Table 2, Table 3 and Table 4 to report the certainty of evidence (GRADEpro GDT).

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

1. High certainty: further research is very unlikely to change our confidence in the estimate of effect.

2. Moderate certainty: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.

3. Low certainty: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

4. Very low certainty: we are very uncertain about the estimate.

Results

Description of studies

Results of the search

We identified 1615 references from database and trial registry searches, and one record from personal contacts (n = 1616). After removing 735 duplicates, 881 titles/abstracts were available for screening.

We excluded 865 records based on title/abstract and reviewed 16 full texts. We excluded 11 studies (Characteristics of excluded studies), and categorised one as 'awaiting classification' (Characteristics of studies awaiting classification). We included four studies in this review (Characteristics of included studies). See Figure 1.

1.

Study flow diagram

Included studies

We included four studies including a total of 138 infants. All four studies compared two different forms of CPAP (Chua 2005; Gutierrez 2003; Pantalitschka 2009; Telenko 1999). Chua 2005 and Gutierrez 2003 were both two‐arm, parallel trials. Pantalitschka 2009 was a four‐period cross‐over study, but only two of the four interventions were relevant to this review. Each period was six hours. Since the total duration of the intervention was 24 hours there appears to be no washout period between each period. The two excluded interventions were both different forms of NIPPV. Telenko 1999 was a three‐period, two‐intervention, double cross‐over study with each period being six hours. Any washout period used was not stated. The study was reported only in abstract form. There were no usable outcome data in the abstract report. Dr Telenko was contacted but was not able to supply any further study information. Two studies were stopped early due to slow recruitment (Chua 2005; Pantalitschka 2009).

The studies were set in Malaysia (Chua 2005), Spain (Gutierrez 2003), Germany (Pantalitschka 2009), and North America (Telenko 1999). One study was not funded (Chua 2005), and the remaining three studies did not report funding.

Participants

The total number of participants was 138 of which 122 had usable data, ranging from 16 to 80 in each study. The mean or median birth weight ranged from 900 g to 1100 g (range 480 g to 1540 g). The participant mean or median gestation was 26 to 28 weeks. The mean or median age at initiation of the intervention was between nine days and three weeks. One study did not report participants' age (Telenko 1999). Three studies excluded infants with risk factors for apnoea such as infection, electrolyte disturbance, and intraventricular haemorrhage. One study included infants with apnoea due to secondary causes (Gutierrez 2003).

Interventions

No study compared CPAP with supportive care, such as oxygen therapy.

All four studies compared one form of CPAP with another. Chua 2005 and Telenko 1999 compared a variable flow device (Infant Flow (Electromedical Equipment) in Chua, and Aladdin Infant Flow (Hamilton Medical) in Telenko) with conventional ventilator CPAP. Pantalitschka 2009 compared underwater bubble CPAP with variable flow CPAP (Infant Flow (Electromedical Equipment)). Gutierrez 2003 compared two forms of variable flow devices (Infant Flow (Electromedical Equipment) with Medijet (Medin)). The CPAP pressures used were between 4 and 6 cm H₂O.

The duration of the intervention in Chua 2005 was 48 hours, six hours in Telenko 1999, six hours in Pantalitschka 2009, and 24 hours in Gutierrez 2003.

For three studies all infants had received methylxanthines (one study each for aminophylline, caffeine, and theophylline), as prophylaxis prior to the initiation of the intervention. One study did not report use of methylxanthines (Telenko 1999).

Outcomes

Two studies reported our primary outcome, failed CPAP (Chua 2005; Gutierrez 2003). In both, the definition for failure was the need for intubation. One study did not state the criteria for intubation and one study stated that the decision was made by clinical judgement of the attending doctor.

The other primary outcome, apnoea, was reported in all four studies. The definitions used included cessation of breathing for 20 seconds in three studies and a reduction in respiratory amplitude of 20% for more than 10 seconds in the fourth study. Definitions of desaturation were a fall to below 80% (two studies), 85%, or 88%. Bradycardia was defined as a fall in heart rate to below 80 or 100 beats per minute (two studies), a reduction of 30% below baseline in one study, and was not defined in one study.

Three studies used data from the monitor memory to calculate apnoea rates. One study did not report how the data were collected. Rates of apnoea were reported in two studies (Chua 2005; Pantalitschka 2009), and one study reported the outcome as resolution of apnoea within 24 hours (Gutierrez 2003).

The secondary outcomes reported were rate of apnoea with bradycardia (Chua 2005), bronchopulmonary dysplasia (oxygen dependency at 36 weeks) (Chua 2005), mortality (Chua 2005), pneumothorax (Chua 2005; Gutierrez 2003; Pantalitschka 2009), and gastrointestinal complications (abdominal distension, abdominal distension requiring cessation of feeds, vomiting) (Chua 2005; Gutierrez 2003).

No study reported long‐term outcomes.

Excluded studies

We excluded 11 studies (Campbell 2006; Claassen 2019; Davis 1998; Davis 2001a; Gross 2020; Heiring 2015; Ishihara 2016; Kattwinkel 1975; NCT00187434; Yu 2016; Zaharie 2008). Four studies included CPAP for weaning from mechanical ventilation (Claassen 2019; Davis 1998; Davis 2001a; NCT00187434). Although some measured AoP as an outcome, this was part of the definition of the outcome failed extubation. Two studies compared CPAP with nasal IPPV including bilevel CPAP (Gross 2020; Ishihara 2016). One study compared CPAP with high flow nasal cannula (Campbell 2006). One study compared methylxanthine and placebo for AoP and both groups received CPAP (Yu 2016). One study compared prophylactic and treatment CPAP for respiratory distress syndrome (Zaharie 2008), and one study compared continuing CPAP versus weaning at 28 days of life (Heiring 2015). Kattwinkel 1975 compared apnoea rates before and after either kinetic stimulation or nasal CPAP. Dr Kattwinkel could not confirm that the patients were randomised to these two interventions.

One further study is awaiting classification (Pantalitskcha 2006). We did not find any ongoing studies.

Risk of bias in included studies

The risk of bias is summarised in Figure 2 and Figure 3.

2.

3.

Allocation

Although all the studies were described as randomised, only one study clearly described adequate random sequence generation (Pantalitschka 2009), and only one study clearly described adequate allocation concealment (Chua 2005). We judged the remaining studies to be unclear.

Blinding

All studies were not blinded to caregivers, and we judged them to be at high risk of bias for blinding of the intervention. Pantalitschka 2009 described blinding of the outcome assessors whereas for Chua 2005 and Gutierrez 2003, both studies that used analysis of monitor recordings for the primary outcome, apnoea, it was unclear whether the outcome assessor was blinded to the intervention. Telenko 1999 did not describe blinding in the abstract report.

Incomplete outcome data

Three studies described complete outcome data (Chua 2005; Gutierrez 2003; Pantalitschka 2009), whereas the abstract report by Telenko 1999 did not give any information.

Selective reporting

One study was prospectively registered, and all prospectively registered outcomes were reported (Pantalitschka 2009). Chua 2005 was not prospectively registered, but all expected outcomes were reported, so we judged this to be at low risk of reporting bias. Gutierrez 2003 reported the primary outcome, apnoea, as resolution at 24 hours. We judged this study to be at unclear risk of reporting bias. Telenko 1999 did not give sufficient information in the abstract report to make a judgement.

Other potential sources of bias

Two studies, both cross‐over studies, did not adequately describe the washout period, and we judged both of these to be at high risk of bias for a carry‐over effect (Pantalitschka 2009; Telenko 1999). In addition, Pantalitschka 2009 appeared to use an analysis method not generally accepted for cross‐over studies.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4

Comparison 1. Continuous positive airway pressure (CPAP) and supportive care compared to supportive care alone

There were no studies comparing CPAP and supportive care with supportive care alone. See Table 1.

Comparison 2. CPAP delivered by different types of devices

Four studies compared one CPAP device with another. Two compared a variable flow device with ventilator CPAP (Chua 2005; Telenko 1999). One compared two different variable flow devices (Gutierrez 2003), and one compared a variable flow device with bubble CPAP (Pantalitschka 2009).

2a. Variable flow compared to ventilator CPAP

See Table 2.

Primary outcomes

Failed CPAP

Only one study reported failed CPAP. We are very uncertain whether there is any difference in the incidence of failed CPAP, defined as the need for mechanical ventilation (risk ratio (RR) 0.16, 95% confidence interval (CI) 0.01 to 2.90; 1 study, 26 participants; very low‐certainty; Analysis 1.1) (Chua 2005).

1.1. Analysis.

Comparison 1: Variable flow versus conventional ventilator CPAP, Outcome 1: Failed CPAP

Apnoea

Chua 2005 reported apnoea events per four‐hour period. We are very uncertain whether there is any difference in the frequency of apnoea events (mean difference (MD) ‐0.10, 95% CI ‐1.30 to 1.10; 1 study, 26 participants; very low‐certainty; Analysis 1.2). Telenko 1999 did not report any analysable data but stated that there was no significant difference in the number of apnoea and bradycardic events per hour of tracing between the two groups.

1.2. Analysis.

Comparison 1: Variable flow versus conventional ventilator CPAP, Outcome 2: Apnoea events

Secondary outcomes

Use of methylxanthines

This outcome was not reported. However, all patients received methylxanthines as part of the intervention.

Rates of apnoea with bradycardia

No study reported this outcome.

Bronchopulmonary dysplasia

One study reported this outcome, defined as oxygen‐dependent at 36 weeks' postmenstrual age (Chua 2005). There were two events, both in the conventional CPAP group. We are very uncertain whether there is any difference in the frequency of bronchopulmonary dysplasia (RR 0.23, 95% CI 0.01 to 4.38; 1 study, 26 participants; very low‐certainty; Analysis 1.3).

1.3. Analysis.

Comparison 1: Variable flow versus conventional ventilator CPAP, Outcome 3: Bronchopulmonary dysplasia at 36 weeks

Mortality

Chua 2005 reported mortality to discharge. There were four deaths, all in the control group (RR 0.13, 95% CI 0.01 to 2.16; 1 study, 26 participants; very low‐certainty; Analysis 1.4).

1.4. Analysis.

Comparison 1: Variable flow versus conventional ventilator CPAP, Outcome 4: Mortality to discharge

Pneumothorax

Chua 2005 reported this outcome. There were no events in either group during the study period.

Retinopathy of prematurity

This outcome was not reported.

Gastrointestinal complications

Chua 2005 reported one episode of abdominal distension in the variable flow group and none in the conventional ventilator CPAP group. In addition, vomiting was not reported in the variable flow group but was seen in two infants in the conventional CPAP group.

Neurodevelopmental complications

No study reported long‐term neurodevelopmental outcomes.

2b. Variable flow compared to bubble CPAP

See Table 3.

Primary outcomes

One study compared a variable flow device with bubble CPAP (Pantalitschka 2009). Of the pre‐specified primary outcomes, only apnoea was reported.

Apnoea

Pantalitschka 2009 reported central apnoeas as medians and range per hour of artefact‐free recording. The medians (range) in the variable flow group and bubble CPAP group were 3.7 (range 1.1 to 26.4) and 6.4 (range 0.9 to 2.41).

The authors also reported the cumulative rate of bradycardias and desaturations with interquartile range (IQR) per hour of artefact‐free recording. The medians (IQR) for the two groups were 2.8 (1.5 to 7.7) and 5.4 (3.0 to 9.8).

Secondary outcomes

Of the pre‐specified secondary outcomes, the only one reported was pneumothorax. The authors stated that there were no adverse events, such as pneumothorax.

2c. Infant Flow compared to Medijet CPAP

See Table 4.

Gutierrez 2003 compared the Infant Flow (Electromedical Equipment) device with the Medijet (Medin) device.

Failed CPAP

Gutierrez 2003 reported this outcome. There were 11 events in the Infant Flow group and four events in the Medijet group. We are very uncertain whether there is any difference in the incidence of failed CPAP (RR 2.62, 95% CI 0.91 to 7.53; 1 study, 80 participants; very low‐certainty, Analysis 2.1).

2.1. Analysis.

Comparison 2: Infant Flow variable flow CPAP versus Medijet variable flow CPAP, Outcome 1: Failed CPAP

Apnoea

Gutierrez 2003 did not report frequency of apnoea but reported number with resolution of apnoea, bradycardia, and desaturations (each as a separate outcome) within 24 hours. There were no differences in any of these outcomes.

Other outcomes

Of the pre‐specified secondary outcomes, Gutierrez 2003 reported only gastrointestinal complications. They reported two participants in each group with abdominal distension.

Comparison 3. CPAP compared to mechanical ventilation

We did not identify any studies with this comparison.

Discussion

Summary of main results

There is no information from randomised controlled trials (RCTs) on the effect of continuous positive airway pressure (CPAP) compared with supportive care on apnoea of prematurity (AoP). All four of the small studies we identified compared CPAP delivered by a variable flow device with another device (conventional ventilator CPAP (two studies), bubble CPAP (one study), and two different variable flow devices (one study)). From these four studies we are very uncertain whether any device is more effective than its comparator. There are no studies comparing CPAP with mechanical ventilation.

Overall completeness and applicability of evidence

We identified three of the four included studies by database searching and one study from other sources. This was an abstract of a thesis, the full text of which we obtained. One study had no funding and three studies did not state the source of funding, so we do not know whether there were any industry‐funded studies included. However, we have no reason to believe that if industry funding was involved, the funders had any influence on the results. We do not expect there to be other unpublished industry‐funded trials that are not available to us.

Given that one study, a thesis, was identified from other sources, there is a possibility that there may be other unidentified theses. Therefore, although we carried out a comprehensive search, which, in addition to database searching, included discussion with experts in the field and examining the reference lists of included and excluded studies, as well as other related systematic reviews, there is a possibility that other unidentified studies, particularly theses, exist.

The included studies were all more than 10 years old, suggesting that this research question is no longer current. Clinical practice guidelines recommend CPAP for apnoea of prematurity (Eichenwald 2016; NICE 2019), and therefore ethics boards may be unwilling to approve further studies, particularly comparing CPAP with supportive care. Nevertheless, this review shows that the evidence from randomised controlled trials is lacking.

The devices included in these studies included two types of variable flow devices, conventional bubble CPAP and conventional ventilator CPAP. For each comparison, data were only available from a single study. There are theoretical reasons why these devices might have different effects on AoP, therefore this review shows that evidence is still lacking on whether any device is more effective than any other. There were no studies for two of our comparisons: CPAP compared with supportive care and CPAP compared with mechanical ventilation.

A number of outcomes could not be reported. We received correspondence from two authors and neither could provide any further information.

We did not attempt to look at the effect of different levels of CPAP pressure on AoP. Another Cochrane Review has looked at different CPAP levels and included apnoea as a secondary outcome, but none of the included studies reported this outcome (Bamat 2021).

Quality of the evidence

We judged the evidence for all the available outcomes as very low‐certainty. The reasons for downgrading included risk of bias (lack of blinding and in some comparisons cross‐over studies with potential carry‐over effect), imprecision due to data being taken from a single small study and wide confidence intervals incorporating both a benefit and a harm. We did not find any evidence of publication bias.

Potential biases in the review process

One potential bias in the review process was that one review author (JJH) was involved in one of the included studies, as the supervisor of the study author's thesis (Chua 2005). JJH was not involved in the data extraction for this study.

In addition, another bias could have arisen when we excluded one study because we could not confirm whether it was an RCT. This was the only study that would have provided information for Comparison 1, CPAP versus supportive care (Kattwinkel 1975).

In one study we found a discrepancy between the results reported in the English abstract and the main text, which was in Spanish. We used the data from the main text. Some of the results were reported as percentages, so we derived absolute numbers from these (Gutierrez 2003).

We also included one cross‐over study that may have had no washout period. We had planned to exclude cross‐over studies with washout periods of less than six hours. However, rather than exclude the study we judged it to be at high risk of bias (Pantalitschka 2009).

Agreements and disagreements with other studies or reviews

We could not find any other systematic reviews assessing the effect of CPAP on AoP. A Cochrane Review evaluating nasal CPAP for extubation included AoP as part of their primary outcome, extubation failure, but did not report apnoea as a separate outcome; in addition, none of the included studies in this review reported apnoea as a separate outcome (Davis 2003). Two systematic reviews have compared nasal intermittent positive pressure (NIPPV) with nasal CPAP for AoP. A Cochrane Review included two small RCTs and the authors were not able to draw any conclusions from these about the effects of the two interventions on AoP (Lemyre 2017). A further systematic review included 18 interventional studies and found no difference in the frequency of apnoea, although a post hoc analysis of 10 studies that reported the incidence of any apnoea resulted in a reduction of this with NIPPV (Sabasi 2021)

Authors' conclusions

Implications for practice.

There are no randomised controlled trials (RCTs) comparing continuous positive airway pressure (CPAP) with other supportive methods such as kinaesthetic stimulation or oxygen therapy for apnoea of prematurity (AoP). As CPAP is already endorsed in clinical guidelines it is unlikely that any RCTs will be conducted to answer this question.

Due to the limited available evidence from the four small studies included in this review, we are very uncertain whether any CPAP device is more effective than its comparator for the prevention and treatment of AoP. The devices included in these studies included two types of variable flow device, conventional bubble CPAP, and conventional ventilator CPAP, and for each comparison data were only available from a single study.

Implications for research.

Although there are no trials comparing CPAP with kinaesthetic stimulation, oxygen therapy, or other supportive care, future trials are unlikely. This is because CPAP for AoP is endorsed in current clinical guidelines and is widely used. Nevertheless, further trials are indicated, particularly in low‐income settings where CPAP devices may not always be readily available.

Since there are theoretical reasons why different CPAP devices might differ in their effects on AoP, further studies are justified in comparing different devices, particularly in extremely preterm infants who may be more likely to fail CPAP treatment. In low‐income settings, trials testing different technologies of low‐cost bubble CPAP against each other, as well as with variable flow devices, might be useful.

We do not expect any further studies comparing CPAP with mechanical ventilation. We believe that this question is historical since mechanical ventilation is no longer standard management for apnoea of prematurity.

The definition of AoP, although somewhat arbitrary, is now fairly widely accepted; future studies should therefore continue to use this definition, unless new further definitions become available.

There are several ways of measuring apnoea. These include cardiorespiratory monitors, video cameras, nasal airflow devices, saturation monitors, impedance devices, or a combination of these. Since there is no standardised method, any method used in future trials needs to be applicable to clinical care and should be carefully described. In addition, the criteria for intervention, including the criteria for initiating CPAP and the criteria for initiating further treatment when CPAP fails, vary widely and therefore also need to be clearly and adequately described.

Although blinding of the intervention is impractical, adherence to strict definitions for failed treatment would help reduce the impact of lack of blinding. Follow‐up studies looking at neurodevelopmental outcomes would be desirable. It would also be helpful if studies were done to determine the validity of methods used for the measurement of AoP, as well as the development of standard criteria for intervention.

What's new

Date	Event	Description
14 September 2023	Amended	Corrected inconsistencies between Results of Search and PRISMA diagram.

History

Protocol first published: Issue 6, 2020
Review first published: Issue 7, 2023

Appendices

Appendix 1. Cochrane Library strategy

	Cochrane Library Issue 9, 2022
	Date Run: September‐06‐2022
ID	Search	Hits
#1	MeSH descriptor: [Infant, Newborn] explode all trees	17651
#2	MeSH descriptor: [Intensive Care, Neonatal] explode all trees	353
#3	MeSH descriptor: [Intensive Care Units, Neonatal] this term only	867
#4	MeSH descriptor: [Gestational Age] this term only	2784
#5	("babe" or "babes" or baby* or "babies" or "gestational age" or "gestational ages" or infant? or "infantile" or infancy or "low birth weight" OR "low birth weights" or "low birthweight" or "low birthweights" or neonat* or "neo‐nat*" or newborn* or "new born?" or "newly born" or "premature" or "pre‐mature" or "pre‐matures" or prematures or prematurity or "pre‐maturity" or "preterm" or "preterms" or "pre term?" or "preemie" or "preemies" or "premies" or "premie" or "VLBW" or "VLBWI" or "VLBW‐I" or "VLBWs" or "LBW" or "LBWI" or "LBWs" or "ELBW" or "ELBWI" or "ELBWs" or "NICU" or "NICUs"):ti,ab,kw OR (("37" or "36" or "35" or "34" or "33" or "32" or "31" or "30" or "29" or "28" or "27" or "26") NEAR/1 (week? or wk?) NEAR/5 (birth? or age? or gestation* or pregnan*)):ti,ab,kw	101399
#6	#1 OR #2 OR #3 OR #4 OR #5	101399
#7	MeSH descriptor: [Positive‐Pressure Respiration] explode all trees	3048
#8	(apvr? or bipap? or bi‐pap?or cpap? or ncpap? or n‐cpap?):ti,ab,kw	1193
#9	((bilevel* or biphasic* or continuous*) NEAR/3 positive airway? pressure?):ti,ab,kw	4730
#10	airway? pressure? release ventilation?:ti,ab,kw	194
#11	((positive pressure? or positive‐pressure?) NEAR/2 (breathing or end expiratory or end‐expiratory or ventilation*)):ti,ab,kw	10139
#12	(ippb? or ippv?):ti,ab,kw	254
#13	#7 OR #8 OR #9 OR #10 OR #11 OR #12	13600
#14	MeSH descriptor: [Apnea] explode all trees	3433
#15	(apnea* or apnoea* or apneic* or apnoeic*):ti,ab,kw	10714
#16	Sleep Hypopne*:ti,ab,kw	2666
#17	(Hypersomnia* OR Periodic Respirat*):ti,ab,kw	621
#18	(sleep disorder* Near/2 breath*):ti,ab,kw OR Aop:ti,ab,kw	3568
#19	#14 OR #15 OR #16 OR #17 OR #18	11808
#20	#6 AND #13 AND #19 [346 trials; 20 reviews]	366

Appendix 2. MEDLINE strategy

	Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process, In‐Data‐Review & Other Non‐Indexed Citations, Daily and Versions 1946 to September 02, 2022
#	Searches	Results
1	Continuous Positive Airway Pressure/	8767
2	(apvr? or bipap? or bi‐pap?or cpap? or ncpap? or n‐cpap?).ti,ab,kw,kf.	2092
3	((bilevel* or biphasic* or continuous*) adj3 positive airway? pressure?).ti,ab,kw,kf.	11867
4	airway? pressure? release ventilation?.ti,ab,kw,kf.	313
5	positive‐pressure respiration/ or intermittent positive‐pressure breathing/ or intermittent positive‐pressure ventilation/	20172
6	((positive pressure? or positive‐pressure?) adj2 (breathing or end expiratory or end‐expiratory or ventilation*)).ti,ab,kw,kf.	7280
7	(ippb? or ippv?).ti,ab,kw,kf.	1063
8	or/1‐7 [CPAP and broader types of ventilation]	36245
9	exp apnea/ or sleep apnea syndromes/ or sleep apnea, central/ or sleep apnea, obstructive/	48710
10	(apnea* or apnoea* or apneic* or apnoeic*).ti,ab,kw,kf.	60052
11	Sleep Hypopne*.ti,ab,kw,kf.	41
12	(Hypersomnia* or Periodic Respirat*).ti,ab,kw,kf.	2357
13	(sleep disorder* adj2 (breath* or respirat*)).ti,ab,kw,kf.	8225
14	AoP.ti,ab,kw,kf.	2224
15	or/9‐14 [Apnea]	75039
16	exp Infant, Newborn/ or Intensive Care, Neonatal/ or Intensive Care Units, Neonatal/ or Gestational Age/	708533
17	(babe or babes or baby* or babies or gestational age? or infant? or infantile or infancy or low birth weight or low birthweight or neonat* or neo‐nat* or newborn* or new born? or newly born or premature or pre‐mature or pre‐matures or prematures or prematurity or pre‐maturity or preterm or preterms or pre term? or preemie or preemies or premies or premie or VLBW or VLBWI or VLBW‐I or VLBWs or LBW or LBWI or LBWs or ELBW or ELBWI or ELBWs or NICU or NICUs).ti,ab,kw,kf.	1008367
18	(("37" or "36" or "35" or "34" or "33" or "32" or "31" or "30" or "29" or "28" or "27" or "26") adj1 (week? or wk?) adj5 (birth? or age? or gestation* or pregnan*)).ti,ab,kw,kf.	55756
19	or/16‐18 [Filter: Neonatal Population 06‐2022‐MEDLINE]	1328886
20	randomized controlled trial.pt.	576279
21	controlled clinical trial.pt.	95017
22	randomized.ti,ab.	623508
23	placebo.ti,ab.	237630
24	drug therapy.fs.	2525293
25	randomly.ti,ab.	391454
26	trial.ti,ab.	714057
27	groups.ti,ab.	2430139
28	or/20‐27 [Cochrane HSSS‐SM Filter; Box 6.4.a Cochrane Handbook]	5516426
29	(quasirandom* or quasi‐random* or randomi* or randomly).ti,ab,kw,kf.	1064982
30	(control* adj2 (group? or random* or trial? or study)).ti,ab,kw,kf.	1061081
31	or/29‐30 [Additional terms to increase sensitivity]	1646799
32	exp animals/ not humans/	5042115
33	(or/28,31) not 32 [RCT Filter: Medline]	5058590
34	meta‐analysis/ or "systematic review"/ or network meta‐analysis/ [/ finds same as.pt. syntax]	285079
35	((systematic* adj3 (review* or overview*)) or (methodologic* adj3 (review* or overview*))).ti,ab,kf,kw.	283284
36	((integrative adj3 (review* or overview*)) or (collaborative adj3 (review* or overview*)) or (pool* adj3 analy*)).ti,ab,kf,kw.	35701
37	(data synthes* or data extraction* or data abstraction*).ti,ab,kf,kw.	36641
38	(hand search* or handsearch*).ti,ab,kf,kw.	10656
39	(mantel haenszel or peto or der simonian or dersimonian or fixed effect* or latin square*).ti,ab,kf,kw.	33028
40	meta‐analysis as topic/ or network meta‐analysis/	25554
41	(meta analy* or metanaly* or meta regression* or metaregression*).ti,ab,kf,kw.	246987
42	(medline or cochrane or pubmed or medlars or embase or cinahl).ab.	301508
43	(cochrane or systematic review?).jw.	19468
44	or/34‐43 [SR filter‐Medline; based on CADTHhttps://www.cadth.ca/strings‐attached‐cadths‐database‐search‐filters]	569500
45	8 and 15 and 19 and 33 [RCT Results]	249
46	8 and 15 and 19 and 44 and 202*.yr. [SR Results]	9
47	or/45‐46	252

Appendix 3. Embase strategy

	Embase 1974 to 2022 September 02
#	Searches	Results
1	Continuous Positive Airway Pressure/	5186
2	(apvr? or bipap? or bi‐pap?or cpap? or ncpap? or n‐cpap?).ti,ab,kw,kf.	4238
3	((bilevel* or biphasic* or continuous*) adj3 positive airway? pressure?).ti,ab,kw,kf.	17500
4	airway? pressure? release ventilation?.ti,ab,kw,kf.	561
5	exp positive‐pressure respiration/	15426
6	((positive pressure? or positive‐pressure?) adj2 (breathing or end expiratory or end‐expiratory or ventilation*)).ti,ab,kw,kf.	10319
7	(ippb? or ippv?).ti,ab,kw,kf.	1461
8	or/1‐7 [CPAP and broader types of ventilation]	39669
9	exp apnea/	29852
10	(apnea* or apnoea* or apneic* or apnoeic*).ti,ab,kw,kf.	97047
11	Sleep Hypopne*.ti,ab,kw,kf.	66
12	(Hypersomnia* or Periodic Respirat*).ti,ab,kw,kf.	3946
13	(sleep disorder* adj2 (breath* or respirat*)).ti,ab,kw,kf.	14240
14	AoP.ti,ab,kw,kf.	3021
15	or/9‐14 [Apnea]	117159
16	newborn apnea/ or newborn apnea attack/	1104
17	newborn/ or prematurity/ or newborn intensive care/ or newborn care/ or gestational age/	755596
18	(babe or babes or baby* or babies or gestational age? or infant? or infantile or infancy or low birth weight or low birthweight or neonat* or neo‐nat* or newborn* or new born? or newly born or premature or pre‐mature or pre‐matures or prematures or prematurity or pre‐maturity or preterm or preterms or pre term? or preemie or preemies or premies or premie or VLBW or VLBWI or VLBW‐I or VLBWs or LBW or LBWI or LBWs or ELBW or ELBWI or ELBWs or NICU or NICUs).ti,ab,kw,kf.	1190689
19	(("37" or "36" or "35" or "34" or "33" or "32" or "31" or "30" or "29" or "28" or "27" or "26") adj1 (week? or wk?) adj5 (birth? or age? or gestation* or pregnan*)).ti,ab,kw,kf.	81016
20	or/16‐19 [Filter: Neonatal Population 06‐2022‐OVID EMBASE]	1465383
21	Randomized controlled trial/ or Controlled clinical study/	917728
22	random$.ti,ab,kw.	1836338
23	Randomization/	94978
24	placebo.ti,ab,kw.	346748
25	((double or single or doubly or singly) adj (blind or blinded or blindly)).ti,ab,kw.	260603
26	double blind procedure/	198771
27	(controlled adj7 (study or design or trial)).ti,ab,kw.	418124
28	parallel group$1.ti,ab.	30031
29	(crossover or cross over).ti,ab.	117949
30	((assign$ or match or matched or allocation) adj5 (alternate or group$1 or intervention$1 or patient$1 or subject$1 or participant$1)).ti,ab.	387674
31	(open adj label).ti,ab.	99788
32	(quasirandom* or quasi‐random* or randomi* or randomly).ti,ab,kw,kf.	1497588
33	(control* adj2 (group? or random*)).ti,ab,kw,kf.	1219222
34	or/21‐33 [ Terms based on Cochrane Central strategy‐https://www‐cochranelibrary‐com.ezproxy.uvm.edu/central/central‐creation]	3137025
35	(exp animals/ or exp invertebrate/ or animal experiment/ or animal model/ or animal tissue/ or animal cell/ or nonhuman/) and (human/ or normal human/ or human cell/)	24067979
36	exp animals/ or exp invertebrate/ or animal experiment/ or animal model/ or animal tissue/ or animal cell/ or nonhuman/	31038946
37	36 not 35 [Animal Exclusion‐https://community‐cochrane‐org.ezproxy.uvm.edu/sites/default/files/uploads/inline‐files/Embase%20animal%20filter.pdf]	6970967
38	34 not 37 [Filter: RCT‐EMBASE]	2697743
39	meta‐analysis/ or "systematic review"/ or "meta analysis (topic)"/ [EMTREE]	525878
40	((systematic* adj3 (review* or overview*)) or (methodologic* adj3 (review* or overview*))).ti,ab,kw.	343212
41	((integrative adj3 (review* or overview*)) or (collaborative adj3 (review* or overview*)) or (pool* adj3 analy*)).ti,ab,kw.	50122
42	(data synthes* or data extraction* or data abstraction*).ti,ab,kw.	45127
43	(hand search* or handsearch*).ti,ab,kw.	12992
44	(mantel haenszel or peto or der simonian or dersimonian or fixed effect* or latin square*).ti,ab,kw.	43539
45	(meta analy* or metanaly* or meta regression* or metaregression*).ti,ab,kw.	314490
46	(medline or cochrane or pubmed or medlars or embase or cinahl).ab.	379122
47	(cochrane or systematic review?).jn,jx.	30567
48	(overview adj2 reviews).ti.	119
49	or/39‐48 [SR Filter: EMBASE based on CADTH filter: Strings‐attached]	808417
50	8 and 15 and 20 and 38 [CPAP and Apnea and Neonate and RCT]	316
51	8 and 15 and 20 and 49 and "202*".yr. [CPAP and Apnea and Neonate and SR]	27
52	50 or 51 [Results Embase]	333

Appendix 4. CINAHL strategy

	CINAHL Complete (EbscoHost)
	September‐06‐2022
	Search modes ‐ Boolean/Phrase; Search Screen ‐ Advanced Search
#	Query	Results
1	(MH "Continuous Positive Airway Pressure")	5,869
2	( TI (apvr# or bipap# or bi‐pap#or cpap# or ncpap# or n‐cpap#) ) OR ( AB (apvr# or bipap# or bi‐pap#or cpap# or ncpap# or n‐cpap#) )	657
3	( TI ((bilevel* or biphasic* or continuous*) N3 positive airway# pressure#) ) OR ( AB ((bilevel* or biphasic* or continuous*) N3 positive airway# pressure#) )	4,180
4	TI (airway# pressure# release ventilation#) OR AB (airway# pressure# release ventilation#)	190
5	(MH "Positive Pressure Ventilation+")	12,586
6	( TI ((positive pressure# or positive‐pressure#) N2 (breathing or end expiratory or end‐expiratory or ventilation*)) ) OR ( AB ((positive pressure# or positive‐pressure#) N2 (breathing or end expiratory or end‐expiratory or ventilation*)) )	2,258
7	( TI (ippb# or ippv#) ) OR ( AB (ippb# or ippv#) )	237
8	S1 OR S2 OR S3 OR S4 OR S5 OR S6 OR S7	14,811
9	(MH "Apnea+")	18,044
10	( TI (apnea* or apnoea* or apneic* or apnoeic*) ) OR ( AB (apnea* or apnoea* or apneic* or apnoeic*) )	18,427
11	TI Hypopne* OR AB Hypopne*	3,530
12	TI (Hypersomnia* OR Periodic Respirat*) OR AB (Hypersomnia* OR Periodic Respirat*)	500
13	TI (sleep disorder* N2 breath*) OR AB (sleep disorder* N2 breath*)	2,820
14	TI AOP OR AB AOP	225
15	S9 OR S10 OR S11 OR S12 OR S13 OR S14	25,064
16	(MH "Infant, Newborn+") OR (MH "Infant, Large for Gestational Age") OR (MH "Infant, Low Birth Weight+") OR (MH "Infant, Postmature") OR (MH "Infant, Premature") OR (MH "Intensive Care, Neonatal+") OR (MH "Intensive Care Units, Neonatal")	162,192
17	TI (("37" or "36" or "35" or "34" or "33" or "32" or "31" or "30" or "29" or "28" or "27" or "26") N1 (week# or wk#) N5 (birth# or age# or gestation* or pregnan*)) OR AB (("37" or "36" or "35" or "34" or "33" or "32" or "31" or "30" or "29" or "28" or "27" or "26") N1 (week# or wk#) N5 (birth# or age# or gestation* or pregnan*))	18,509
18	TI ((babe or babes or baby* or babies or gestational age# or infant# or infantile or infancy or low birth weight or low birthweight or neonat* or neo‐nat* or newborn* or new born# or newly born or premature or pre‐mature or pre‐matures or prematures or prematurity or pre‐maturity or preterm or preterms or pre term# or preemie or preemies or premies or premie or VLBW or VLBWI or VLBW‐I or VLBWs or LBW or LBWI or LBWs or ELBW or ELBWI or ELBWs or NICU or NICUs) OR AB (babe or babes or baby* or babies or gestational age# or infant# or infantile or infancy or low birth weight or low birthweight or neonat* or neo‐nat* or newborn* or new born# or newly born or premature or pre‐mature or pre‐matures or prematures or prematurity or pre‐maturity or preterm or preterms or pre term# or preemie or preemies or premies or premie or VLBW or VLBWI or VLBW‐I or VLBWs or LBW or LBWI or LBWs or ELBW or ELBWI or ELBWs or NICU or NICUs)	259,578
19	S16 OR S17 OR S18	321,224
20	(MH "Single‐Blind Studies") OR (MH "Triple‐Blind Studies") OR (MH "Randomized Controlled Trials+") OR (MH "Double‐Blind Studies")	172,790
21	(MH "Clinical Trials+")	343,194
22	TI (randomi#ed or randomly) OR AB (randomi#ed or randomly) OR AB (randomi#ed or randomly) OR AB (randomi#ed or randomly)	362,307
23	AB randomly	105,677
24	AB placebo	65,321
25	AB (trial)	333,867
26	AB groups	883,480
27	TI (quasirandom* or quasi‐random*) OR AB (quasirandom* or quasi‐random*)	2,202
28	S20 OR S21 OR S22 OR S24 OR S25 OR S26 OR S27	1,292,625
29	(MH "Animal Studies")	148,816
30	(MH "Human")	2,596,757
31	S29 NOT S30	123,925
32	S28 NOT S31	1,253,377
33	(MH "Systematic Review")	113,269
34	(MH "Meta Analysis")	65,009
35	( TI ((systematic* N3 (review* or overview*)) or (methodologic* N3 (review* or overview*))) ) OR ( AB ((systematic* N3 (review* or overview*)) or (methodologic* N3 (review* or overview*))) )	139,882
36	( TI ((integrative N3 (review* or overview*)) or (collaborative N3 (review* or overview*)) or (pool* N3 analy*)) ) OR ( AB ((integrative N3 (review* or overview*)) or (collaborative N3 (review* or overview*)) or (pool* N3 analy*)) )	18,635
37	( TI (data synthes* or data extraction* or data abstraction*) ) OR ( AB (data synthes* or data extraction* or data abstraction*) )	13,999
38	AB (hand search* or handsearch*)	4,897
39	AB (mantel haenszel or peto or der simonian or dersimonian or fixed effect* or latin square*)	9,508
40	( TI met analy* or metanaly* or meta regression* or metaregression*) ) OR ( AB met analy* or metanaly* or meta regression* or metaregression*) )	4,873
41	AB (medline or cochrane or pubmed or medlars or embase OR CINAHL)	116,856
42	S33 OR S34 OR S35 OR S36 OR S37 OR S38 OR S39 OR S40 OR S41 [Published Date: 20200101‐20220931 ]	66,406
43	S8 AND S15 AND S19	219
44	S8 AND S15 AND S19 AND S32	83
45	S8 AND S15 AND S19 AND S42	1
46	S44 OR S45	84

Appendix 5. Trial registry strategies

	ICTRP
September‐06‐2022	cpap [Title]; Apnea [Condition]; trials in children	2
September‐06‐2022	continous positive airway [Title]; Apnea [Condition]; trials in children	1
September‐06‐2022	continous positive airway [intervention]; Apnea [Condition]; trials in children	3
	ISRCTN
	cpap within Condition Category: Neonatal Diseases AND Condition: Apnea	1
June‐15‐2022	Interventions: Cpap Remove filter Condition: Apnea Remove filter	0
June‐15‐2022	Interventions: Continuous positive Remove filter Condition: Apnea Remove filter	0
June‐15‐2022	Condition: Apnea ; Age Range Neonate	1
	Clinicaltrials.gov
September‐06‐2022	Condition: Apnea; Other terms: Cpap. Age range Child birth ‐ 17	74
September‐06‐2022	Condition: Apnea; Other terms: continuous positive airway pressure Age range Child birth ‐ 17	56
June‐15‐2022	Condition: Apnea; Other terms: continuous positive airway pressure neonate	13
June‐15‐2022	Condition: Apnea; Other terms: Ncpap neonate	4
June‐15‐2022	Condition: apnea Other terms: cpap neonate	14
June‐15‐2022	Condition: Apnea; Other terms: premature continuous positive airway	23
June‐15‐2022	Codition: Apnea; Other terms premature cpap	29
September‐06‐2022	Other terms: apnea cpap neonate ; also tried Newborn and retrieved same 65	65
September‐06‐2022	Other terms: apnea cpap premature	60
June‐15‐2022	Other terms: apnea continuous positive airway pressure neonate; tried also newborn and got same 51	51
	ICTRP
June‐15‐2022	apnea cpap; clinical trials in children	7
June‐15‐2022	Title: newborn; Condition: Apnea Intervention CPAP	0
June‐15‐2022	Title: newborn; Condition: Apnea Intervention Continuous positive airwary pressure	0
June‐15‐2022	Title: neonate; Condition: Apnea Intervention Continuous positive airwary pressure
June‐15‐2022	Main search box: apnea continuous positive airway pressure; Trials in children	20
June‐15‐2022	Main search box: apnea continuous positive airway pressure neonate	0
June‐15‐2022	Main search box: cpap neonate	11
June‐15‐2022	Main search box: cpap newborn	75
June‐15‐2022	Main search box: apnea neonate	7
June‐15‐2022	Main search box: apnea newborn	24
June‐15‐2022	Main search box: ncpap newborn	32
June‐15‐2022	Main search box: ncpap neonate	7
	Total	580

Appendix 6. Risk of bias tool

We used the standard methods of Cochrane and Cochrane Neonatal to assess the methodological quality of the trials. For each trial, we sought information regarding the method of randomisation, blinding, and reporting of all outcomes of all infants enrolled in the trial. We assessed each criterion as being at low, high, or unclear risk of bias. Two review authors separately assessed each study. Any disagreements were resolved by discussion. We added this information to the Characteristics of included studies table. We evaluated the following issues and entered the findings into the risk of bias table.

1. 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.

2. 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

3. 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. We assessed blinding separately for different outcomes or classes 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.

4. 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. We assessed blinding separately for different outcomes or classes of outcomes. We categorised the methods as:

• low risk for outcome assessors;

• high risk for outcome assessors; or

• unclear risk for outcome assessors.

5. 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 when reported, and whether missing data were balanced across groups or were related to outcomes. When 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 (< 20% missing data);

• high risk (≥ 20% missing data); or

• unclear risk.

6. Selective reporting bias. Are reports of the study free of the 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 pre‐specified 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 (when it is clear that all of the study's pre‐specified outcomes and all expected outcomes of interest to the review have been reported);

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

• unclear risk.

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

For each included study, we described any important concerns we had about other possible sources of bias (e.g. whether there was a potential source of bias related to the specific study design, 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 planned to explore the impact of the level of bias by undertaking sensitivity analyses.

Data and analyses

Comparison 1. Variable flow versus conventional ventilator CPAP.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Failed CPAP	1	26	Risk Ratio (M‐H, Fixed, 95% CI)	0.16 [0.01, 2.90]
1.2 Apnoea events	1	26	Mean Difference (IV, Fixed, 95% CI)	‐0.10 [‐1.30, 1.10]
1.3 Bronchopulmonary dysplasia at 36 weeks	1	26	Risk Ratio (M‐H, Fixed, 95% CI)	0.23 [0.01, 4.38]
1.4 Mortality to discharge	1	26	Risk Ratio (M‐H, Fixed, 95% CI)	0.13 [0.01, 2.16]
1.5 Gastrointestinal complications	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
1.5.1 Abdominal distension	1	62	Risk Ratio (M‐H, Fixed, 95% CI)	2.19 [0.09, 51.70]
1.5.2 Vomiting	1	52	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 7.82]

1.5. Analysis.

Comparison 1: Variable flow versus conventional ventilator CPAP, Outcome 5: Gastrointestinal complications

Comparison 2. Infant Flow variable flow CPAP versus Medijet variable flow CPAP.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Failed CPAP	1	80	Risk Ratio (IV, Fixed, 95% CI)	2.62 [0.91, 7.53]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Chua 2005.

Study characteristics
Methods	Randomised controlled trial
Participants	Inclusion criteria: all preterm babies ≤ 32 weeks at birth and corrected age ≤ 34 weeks at the time of recruitment with apnoea of prematurity (at least 2 episodes over 6 hours) with secondary causes excluded and required CPAP to control the apnoea as judged by the attending medical officer Experimental group: n = 12, median (IQR) gestation 28 weeks (1), birth weight 1102 g (183), weight at recruitment 1085 g (373), age at recruitment 9.5 days (12), male sex 25% Control group: n = 14, median (IQR) gestation 28 weeks (2), birth weight 1051 g (182), weight at recruitment 1030 g (255), age at recruitment 8.5 days (9), male sex 64% Imbalance for males (P = 0.045) No significant differences in Hb, HCT, sodium, potassium, calcium, random blood sugar Study set in tertiary NICU in Malaysia. Study dates: December 2003 to April 2005.
Interventions	1. Experimental: nasal Infant Flow driver CPAP Level of CPAP: 5 cm H2O Manufacturer of device: Electro Medical Equipment Ltd, UK Interface: nasal prongs with chin strap 2. Control: ventilator CPAP using 1 of 5 conventional ventilators: Sechrist IV‐100B, Bear Cub BP 2001, Millennium Sechrist Level of CPAP: 5 cm H2O, flow rate 6 to 8 L/min Interface: nasal prongs (Argyle) with chin strap In both groups the FiO2 was adjusted to maintain SaO2 at 89% to 95%. Duration of intervention period was 48 hours. All infants recruited received aminophylline loading dose 10 mg/kg/dose at 1 day of life followed by a maintenance dose of 2.5 mg/kg/dose daily
Outcomes	Prespecified: Failed CPAP – need for intubation and additional respiratory support, judged by the attending doctor. No criteria were used (would have resulted in a change in practice in the unit). Total rate of apnoea expressed as events per 4‐hour period Rate of moderate and severe apnoea expressed as events per 4‐hour period Oxygen‐dependent at 36 weeks corrected age Death before hospital discharge Pneumothorax Gastrointestinal complications (abdominal distension, abdominal distension requiring cessation of feeds, number of episodes of vomiting, grading up of feeds during the 48‐hour study period) Apnoea was also graded into mild (cessation of breathing for more than 20 seconds without bradycardia or desaturation), moderate (cessation of breathing (> or < 20 seconds) with either bradycardia or desaturation), and severe (cessation of breathing (> or < 20 seconds) with both bradycardia or desaturation) Other outcomes reported: Combined outcome of failed CPAP and death before discharge Nasal injury FiO2 requirements before and 48 hours after the study Grading of feeds during the 48 study period
Notes	Abstract was published as conference proceedings. Unpublished. Full report only available as an unpublished thesis obtained from the author. Did not meet the calculated sample size due to slow recruitment. No funding. Competing interest: the study author's thesis supervisor is a review author (JJH). However, JJH was not involved in the data extraction of this study.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "randomly assigned"; method of randomisation not described
Allocation concealment (selection bias)	Low risk	Quote: "sealed sequentially numbered opaque envelope"
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "... there was no blinding of the intervention". The primary outcome failed treatment was not based on pre‐specified criteria. Judgement was made by the attending physician.
Blinding of outcome assessment (detection bias) All outcomes	High risk	The decision for failed treatment was made by the attending physician using pre‐specified criteria but these were not very specific and open to judgement. The primary outcome apnoea frequency was taken from the monitor memory but would involve decisions by the attending physician. Both outcomes judged to be at high risk of detection bias.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts. All infants accounted for.
Selective reporting (reporting bias)	Low risk	No protocol available. All pre‐specified primary and secondary outcomes were reported. All expected outcomes were reported.
Other bias	Low risk	Even though there was a baseline imbalance for infant sex (significantly more boys), this appears to be a chance imbalance. Boys may have more severe illness than girls, but as the authors pointed out of the 4 deaths that occurred during the study, 3 were girls.

Gutierrez 2003.

Study characteristics
Methods	Parallel randomised controlled trial
Participants	Preterm infants that required nasal CPAP. Group III: apnoea‐bradycardia syndrome with methylxanthine resistance (previous theophylline between 5 and 12 g/mL) and presence of hypoxic episodes presenting as apnoeas, bradycardias, or desaturations (5 to 7 episodes in a 24‐hour period, or 1 episode per hour over 2 to 3 hours, or more than 2 episodes in the first 24 hours post‐extubation or 1 that requires manual ventilation support). Apnoea was considered a pause for more than 20 seconds and bradycardia a heart rate below 100/minute, and desaturation below 80% on the monitor Group 1: (n = 41), birth weight: 1.031 ± 338, gestation: 27.6 ± 2.1, apnoea aetiology: primary or idiopathic (80%), infectious (20%), neurologic (0%) Group 2: (n = 39), birth weight: 972 ± 321, gestation: 26.8 ± 2.5, apnoea aetiology: primary or idiopathic (56%), infectious (36%), neurologic (8%) Setting: NICU in Spain. Study dates March 2000 to March 2001.
Interventions	Group 1: nasal CPAP delivered by Infant Flow device (a variable flow device creating a Coanda effect) Level of CPAP pressure: 4 to 6 cm H2O Manufacturer of device: Electromedical Equipment Type of nasal interface: nasal prong Group 2: nasal CPAP delivered by Medijet (a variable flow device using a modified Benveniste valve) Level of CPAP pressure: 4 to 6 cm H2O Manufacturer of device: Medin Type of nasal interface: nasal prong The FiO2 was adjusted to maintain saturation between 89% and 93%. Flow was 5 to 8 L/min. All infants were on theophylline.
Outcomes	Failed therapy (need for intubation within 48 hours) Resolution of apnoea within 24 hours (categorical outcome), resolution of bradycardia, and resolution of desaturations each reported separately Mean duration of treatment (hours and range) Pneumothorax* Gastrointestinal complications (abdominal distension)* *These 2 outcomes were not reported separately for the 2 participant groups but included all infants irrespective of the indication for CPAP (see note below)
Notes	The trial report included 3 groups of infants (infants with RDS, infants ready for extubation from mechanical ventilation, and infants with apnoea). Each group were randomly assigned to 1 or other of the 2 CPAP systems. We have included only the group who required CPAP for apnoea. The outcomes pneumothorax and abdominal distension were not reported for each intervention but for all 3 participant groups. The article was published in Spanish. We sought the help of 2 translators who independently extracted the data and where there was disagreement, 1 of the review authors used Google Translate to resolved differences. All outcomes were reported as percentages. Absolute numbers were derived from these using an assumed denominator. We assumed that all infants completed the study. Competing interests not stated. Funding source not stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Prospective randomised trial"; method of random sequence generation not described.
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not described but unlikely due to the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	High risk	Blinding not described but unlikely. Primary outcome subjective outcome.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts, all infants analysed.
Selective reporting (reporting bias)	Unclear risk	No protocol. Frequency of apnoea not reported. Reported as resolution within 24 hours.
Other bias	Low risk	None detected

Pantalitschka 2009.

Study characteristics
Methods	Randomised, controlled cross‐over trial with 4 different interventions. 2 of the 4 interventions are included in this review and 2 were judged not relevant. Each period was 6 hours. There was no washout period described but assumed none because the total duration of intervention was described as 24 hours.
Participants	1) Gestational age at birth < 34 weeks; 2) postmenstrual age and body weight at time of study ≤ 38 weeks and > 1000 g respectively; 3) a requirement for NCPAP to treat AoP as judged by the attending neonatologist. Not requiring oxygen. Excluded: infants with congenital or chromosomal abnormalities, acute infections, intraventricular haemorrhage, requirement for additional inspired oxygen, or patent ductus arteriosus Baseline characteristics (all medians and ranges): (n = 16, 10 boys, 6 girls), median birth weight (g) 1013 (range 480 to 1360), gestational age at birth (weeks) 28 (24 to 29), body weight at time of study (g) 1275 (1030 to 1740), age at time of study (days) 18 (3 to 70), corrected gestational age at time of study (weeks) 31 (28 to 34)
Interventions	Group 1. Variable flow NCPAP device 5 to 6 cm H2O (Infant Flow drive, Electromedical Equipment) Level of CPAP: 5 to 6 cm H2O Manufacturer of device: Electromedical Equipment Interface: Hudson prongs Group 2. Conventional underwater bubble CPAP Level of CPAP: 5 to 6 cm H2O; flow adjusted to keep pressure constant (actual flow rate used not stated) Manufacturer: conventional locally assembled Interface: Hudson prong Group 3: Conventional ventilator delivering NIPPV via Hudson prongs Group 4: NIPPV using a variable flow device (Infant Flow Advance) Each device was applied for 6 hours, yielding a total study duration of 24 hours for each patient All patients were on caffeine Group 1 and 2 included in this review
Outcomes	Cumulative event rate of all bradycardias (< 80 beats per minute) and desaturation events per hour (< 80%) of artefact‐free recording time Rate of central apnoeas per hour Rate of desaturation events per hour Rate of bradycardias per hour All outcomes except respiratory rate measured as event rates per artefact‐free hour of recording using a computerised polygraphic system (Embla N7000 and Somnologica Studio 3.0; Embla, Broomfield, CO) Most outcomes presented as medians and ranges. IQR reported for some outcomes. Where means are presented, the variation is expressed as a range. No SD; SE for MD is given. Although text states that pair‐wise analysis is performed exact P values or t‐statistics are not given for the relevant comparisons. Statistical adjustment for period was made. Data therefore could not be meta‐analysed. Other outcomes reported: Baseline heart rate, baseline respiratory rate and baseline saturation (%). Cumulative event time: the summed duration of all bradycardias and desaturation events divided by artefact‐free recording time and multiplied by 100.
Notes	Study failed to reach calculated sample size due to slow recruitment. Set in NICU in Germany between June 2004 and January 2006. Competing interests declared as none. Funding source not stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Following recruitment, infants were allocated to a random sequence of four different NCPAP/NIPPV devices. The random sequence…was created by Byers’ random selection algorithm."
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not reported but due to the nature of the intervention clinical staff would have been aware of the intervention group.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: “Recordings were analysed without access to clinical data, that is, the author (JS) who analysed the recordings was not involved in clinical management, and analysis of cardio‐respiratory events was carried out without access to the video frame to ensure blinding to the NCPAP/NIPPV devices used.”
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts.
Selective reporting (reporting bias)	Low risk	Trial prospectively registered. All pre‐specified outcomes reported.
Other bias	High risk	Cross‐over study with no washout period and hence potential carry‐over effect. Accepted analysis methods of cross‐over design not used.

Telenko 1999.

Study characteristics
Methods	Randomised, controlled, double cross‐over trial with 36‐hour periods
Participants	Inclusion criteria not given except stated, very low birth weight infants with episodes of apnoea, bradycardia, and desaturations. Baseline characteristics: birth weight 540 g to 1540 g (mean 902 g), gestation 23 to 30 weeks (mean 26.3 weeks), with frequent apnoea, bradycardia, and desaturations. Setting North America. Study dates not reported.
Interventions	Intervention 1. Conventional ventilator CPAP (Infant Star 500 (Infrasonics Inc.)) 6 cm H2O nasal prongs Intervention 2. Variable flow CPAP (Aladdin Infant Flow system (Hamilton Medical)) 6 cm H2O nasal prongs with injector jets
Outcomes	Number of apnoeas, bradycardias, and desaturations per hour of tracing (taken from monitor recordings) Apnoea was defined as cessation of breathing > 20 seconds Bradycardia – a fall in heart rate 30% below baseline Desaturation – a fall in saturation below 88%
Notes	Trial reported in abstract form only. Not published. Authors contacted for further information. They stated that they no longer have access to the trial data. Not registered. No funding statement and no statement of competing interests.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported (Quote:) "... 16 infants were prospectively randomised ..."
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unlikely due to the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated that the recordings were read blinded
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Not stated
Selective reporting (reporting bias)	Unclear risk	None known
Other bias	High risk	Washout period not described but from methods report seems unlikely. Carry‐over effect likely.

AoP: apnoea of prematurity; CPAP: continuous positive airway pressure; FiO₂: fraction of inspired oxygen; HCT: haematocrit; IQR: interquartile range; MD: mean difference; NCPAP: nasal continuous positive airway pressure; NICU: neonatal intensive care unit; NIPPV: nasal intermittent positive airway pressure; RDS: respiratory distress syndrome; S_aO₂: arterial oxygen saturation; SD: standard deviation; SE: standard error

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Campbell 2006	RCT comparing CPAP with high‐flow nasal cannula for infants with respiratory distress after extubation
Claassen 2019	RCT examining 2 different forms of NCPAP for extubation
Davis 1998	RCT comparing CPAP and head box for infants with respiratory distress after extubation
Davis 2001a	RCT comparing 2 modes of nasal CPAP for infants with respiratory distress after extubation
Gross 2020	The 3 interventions included conventional CPAP, variable flow CPAP, and synchronised nasal IPPV. The 2 forms of nasal CPAP were compared as a before and after comparison. Only the variable CPAP and synchronised CPAP were randomly assigned.
Heiring 2015	RCT comparing continuing CPAP with low‐flow nasal cannula in stable infants at 28 postnatal days. Main outcome was lung function. Did not report AoP.
Ishihara 2016	Bilevel CPAP is considered a form of nasal IPPV
Kattwinkel 1975	Comparative study. The author was contacted and could not confirm patients were randomised.
NCT00187434	RCT comparing 2 forms of CPAP for extubation. Trial results not available.
Yu 2016	RCT examining the effect of caffeine for AoP. Both groups received CPAP.
Zaharie 2008	Non‐randomised study examining prophylactic CPAP versus therapeutic CPAP for infants with respiratory distress

AoP: apnoea of prematurity; CPAP: continuous positive airway pressure; IPPV: intermittent positive airway pressure; NCPAP: nasal continuous positive airway pressure; RCT: randomised controlled trial

Characteristics of studies awaiting classification [ordered by study ID]

Pantalitskcha 2006.

Methods	—
Participants	—
Interventions	—
Outcomes	—
Notes	Title only: we have not been able to obtain the abstract

Differences between protocol and review

We made the following changes to the protocol (Ho 2020c).

1. In the protocol we used the term 'usual care' to describe the basic care given for apnoea of prematurity, and in the review we have changed this to 'supportive care'. As in the protocol, supportive care might mean physical stimulation, the administration of oxygen, or both.

2. In the methods section of the review, we have added in the section Unit of analysis issues how we dealt with cross‐over studies with an unclear washout period.

3. We added an additional critical outcome (Long‐term neurodevelopmental outcomes) to the summary of findings tables.

Contributions of authors

JJH conceived the study, and all review authors contributed to the development of the protocol and the review.

IZG, EL, and JWL drafted the protocol, and all review authors commented on the protocol and approved it before publication.

TItle/abstract screening: EL JWL, JJH

Full‐text screening: JJL, IZG. All authors were involved in the final selection of studies.

Data extraction and assessment of risk of bias: JJL, JWL, JJH

Drafting of review: JJH

Commenting on the draft: all authors

Sources of support

Internal sources

• No sources of support provided

External sources

• Vermont Oxford Network, USA

  Vermont Oxford Network is a not‐for‐profit voluntary collaboration of neonatal intensive care units dedicated to improving newborn care throughout the world. Vermont Oxford Network supports the administrative needs of Cochrane Neonatal.

Declarations of interest

JJH is an editor with Cochrane Neonatal. However, she had no involvement in the editorial processing of this review. She was also involved as a thesis supervisor for one of the included studies (Chua 2005).

PS has no conflict of interest to declare.

IZG has no conflict of interest to declare.

JJL has no conflict of interest to declare.

EL has no conflict of interest to declare.

JWL has no conflict of interest to declare.

Edited (no change to conclusions)