Biphasic Positive Airway Pressure or Continuous Positive Airway Pressure: A Randomized Trial
BACKGROUND: There is currently no clear evidence that nasal-biphasic positive airway pressure (n-BiPAP) confers any advantage over nasal-continuous positive airway pressure (n-CPAP). Our hypothesis was that preterm infants born before 30 weeks' gestation and <2 weeks old when extubated onto n-BiPAP will have a lower risk of extubation failure than infants extubated onto n-CPAP at equivalent mean airway pressure.
METHODS: We conducted an unblinded multicenter randomized trial comparing n-CPAP with n-BiPAP in infants born <30 weeks' gestation and <2 weeks old. The primary outcome variable was the rate of extubation failure within 48 hours after the first attempt at extubation. Block randomization stratified by center and gestation (<28 weeks or ≥28 weeks) was performed.
RESULTS: A total of 540 infants (270 in each group) were eligible to be included in the statistical analysis; 57 (21%) of n-BiPAP group and 55 (20%) of n-CPAP group failed extubation at 48 hours postextubation (adjusted odds ratio 1.01; 95% confidence interval 0.65–1.56; P = .97). Subgroup analysis of infants born before and after 28 weeks’ gestation showed no significant differences between the 2 groups. There were no significant differences between arms in death; oxygen requirement at 28 days; oxygen requirement at 36 weeks' corrected gestation; or intraventricular hemorrhage, necrotizing enterocolitis requiring surgery, or pneumothorax.
CONCLUSIONS: This trial shows that there is no added benefit to using n-BIPAP over n-CPAP at equivalent mean airway pressure in preventing extubation failures in infants born before 30 weeks' gestation and <2 weeks old.
- CI —
- confidence interval
- n-BiPAP —
- nasal-biphasic positive airway pressure
- n-CPAP —
- nasal-continuous positive airway pressure
- ITT —
- intention to treat
- MAP —
- mean airway pressure
- PDA —
- patent ductus arteriosus
What’s Known on This Subject:
There have been few published trials to date directly comparing the efficacy of nasal-biphasic positive airway pressure with nasal-continuous positive airway pressure in preterm infants born before 30 weeks’ gestation and <2 weeks old.
What This Study Adds:
This trial demonstrates that nasal-biphasic positive airway pressure confers no significant benefit over nasal-continuous positive airway pressure in preventing extubation failure at equivalent mean airway pressures in preterm infants born before 30 weeks’ gestation and <2 weeks old.
Infants born prematurely before 30 weeks’ gestation are more likely to require ventilator support in the immediate neonatal period due to lung immaturity. Nasal-continuous positive airway pressure (n-CPAP) has been shown to reduce the risk of extubation failure in this group of infants.1 In recent years, nasal-biphasic positive airway pressure (n-BiPAP) has been introduced as an alternative to conventional n-CPAP but clear evidence of its benefit for immediate support after primary extubation compared with n-CPAP is so far lacking. In this study, we aimed to determine if the use of n-BiPAP is more effective than single-level variable-flow n-CPAP in preventing extubation failure when delivered at equivalent mean airway pressure (MAP) in infants born before 30 weeks’ gestation.
An unblinded multicenter randomized trial of n-BIPAP versus n-CPAP in infants born before 30 weeks’ gestation and <2 weeks old was performed with an allocation ratio of 1:1. The trial protocol has been described in greater detail elsewhere.2 No changes to trial protocol occurred after trial commencement. Our specific objective was to conduct a randomized controlled trial in infants born before 30 weeks’ gestation and <2 weeks old to compare the risk of extubation failure over 48 hours after the first attempt at extubation to either n-BiPAP or n-CPAP at equivalent MAP.
The study was conducted in the northwest of England and was approved by North West 11 Research Ethics Committee. Eight regional NICUs participated in the study. Infants born before 30 weeks’ gestation and <2 weeks old were eligible to participate in the study at the time of first extubation attempt after informed parental consent. Parents were approached from soon after birth and well in advance of an extubation attempt. Infants with major congenital malformations, upper respiratory tract abnormalities, and neuromuscular disease were excluded. Infants known to have significant cranial ultrasound scan abnormalities (intraventricular hemorrhage with ventricular dilatation and/or parenchymal extension or white matter abnormalities) before extubation and those likely to be within 7 days’ postlaparotomy at the time of extubation also were excluded.
Interventions and Treatment Strategies
To be eligible for the first extubation attempt, the infant needed to be (1) loaded with caffeine according to standard local protocol and have good respiratory effort persistently higher than the ventilator rate, and (2) have satisfactory blood gases (defined as pH >7.25 and partial pressure of carbon dioxide <7 kPa [52.5 mm Hg]) on minimum ventilation (defined as MAP ≤7 cm water and fractional inspired oxygen concentration of ≤0.35). Unplanned extubation was included if these criteria were satisfied within 4 hours before extubation and informed consent had been obtained.
The n-CPAP group received at extubation a single-level continuous positive airway pressure of 6 cm water for at least 48 hours before weaning was commenced. If the infant was stable for the preceding 24 hours, defined as having <3 minor apneas and no increase in oxygen requirement, weaning was permitted. Minor apnea was defined as apnea requiring stimulation but not mask ventilation. n-CPAP was decreased from 6 cm water by 1 cm water every 48 hours if tolerated based on these criteria. This was done until a pressure of 4 cm water was reached. If a pressure of 4 cm water was successfully tolerated for 48 hours, then time off n-CPAP was allowed. Thereafter, no fixed weaning regimen was prescribed.
The n-BiPAP group received at extubation am MAP of 6 cm water (baseline pressure of 4 cm water and peak pressure of 8 cm of water). Inspiratory time was set at 1 second and rate was set at 30 per minute. If the infant was stable for the preceding 48 hours at MAP of 6 cm water, weaning to MAP of 5 cm water (baseline pressure of 4 cm water and peak pressure of 6 cm of water) was permitted. If the infant was stable for the preceding 48 hours at MAP of 5 cm water, weaning to n-CPAP of 4 cm water was permitted. If the infant was stable for the preceding 48 hours at n-CPAP of 4 cm water, then time off n-CPAP was permitted. Thereafter, no fixed weaning regimen was prescribed.
Routine hourly monitoring of heart rate, peripheral oxygen saturation, and respiratory rate was performed as per standard care. Monitored oxygen saturations were maintained with the range 90% to 95% in all participating NICUs. Delivered n-CPAP and n-BiPAP pressures were monitored regularly and efforts were made to maintain desired n-CPAP and MAP levels in accordance with standard nursing practice. Active mouth closure with chin strap was not performed. Infants were followed until death or discharge home from hospital. Surfactant was administered routinely for all ventilated infants soon after intubation.
Infants were determined to have a failed extubation attempt if there was (1) uncompensated respiratory acidosis defined as pH <7.2 and partial pressure of carbon dioxide >8 kPa (60 mm Hg), or (2) major apnea requiring mask ventilation during the first 7 days postextubation. Crossover was not allowed when the criteria for failure of extubation were reached. Rescue treatment was provided by increasing pressures up to 6 cm of water if weaned or by reintubation and ventilation.
Failure of extubation during the first 48 hours postextubation was the primary outcome measure. Prespecified secondary outcome measures were (1) maintenance of successful extubation for 7 days from the hour of extubation; (2) number of ventilator days after first extubation attempt; (3) oxygen requirement at 28 days of age and at 36 weeks’ corrected gestation; (4) pH, partial pressure of carbon dioxide in the first postextubation gas done within 2 hours after extubation; (5) duration of hospitalization; (6) rates of abdominal distension requiring cessation of feeds for 7 days postextubation; (7) rate of apnea and bradycardia expressed as events per hour during the 48 hours after extubation; and (8) age at transfer back to referral center in days.
A sample size of 270 in each group was planned to give 80% power to detect a 10% reduction in the rates of extubation failure from 25% in the n-CPAP group to 15% in the n-BiPAP group at a 0.05 2-sided significance level. A 10% reduction in extubation failure rate was considered to be clinically significant and could support a change in practice to using n-BiPAP as first line treatment post-extubation.
Infants were randomized afterr the decision to extubate by using Web-based randomization stratified by center and gestation (<28 weeks or ≥28 weeks). The allocation sequence was computer-generated with blocks of random size between 2 and 8.
The participants were enrolled by the clinical team. The randomization was performed by the clinical team just before extubation. Intervention was commenced within 4 hours of the blood gas on which randomization was performed.
The n-BiPAP device produces an audible noise that cannot be masked. Because of this, parents, clinicians involved in patient care, and researchers assessing study endpoints were not blinded to the nature of the study treatments.
A formal statistical analysis plan was prespecified. The primary efficacy analysis was conducted on an intention-to-treat (ITT) basis. As the primary outcome variable is a binary (yes/no) outcome variable, groups were compared by using a logistic regression model adjusting for the stratification variables. Effect sizes are summarized as odds ratios with 95% confidence intervals (CI) and likelihood-ratio–based significance levels computed.
The ITT dataset comprised all correctly randomized patients based on trial inclusion and exclusion criteria. Sensitivity analyses were conducted by using a per-protocol dataset that excluded all participants with major protocol breaches. Major protocol breaches were those that occurred during the randomization process and within 48 hours after randomization and could potentially affect the primary outcome.
A subgroup analysis of the 2 gestation strata was prespecified: an interaction term added to the model and a likelihood-ratio test was used to determine if there was any difference between the gestation groups and odds ratios presented for the 2 subgroups.
Secondary binary outcomes were analyzed by using the same approach and secondary numerical outcomes were analyzed by using similarly adjusted ordinary regression models. Time to event outcomes (days on ventilation and hospital stays) were log-transformed for analysis as log (time+1).
No interim analyses were planned or performed. Analyses were conducted in the R statistical environment v 3.1.0 (R Foundation for Statistical Computing, Vienna, Austria; www.R-project.org/).
Data collection was performed by trained research staff on trial-specific case report forms and entered on a Web-based electronic case record form provided by OpenCDMS.3 Safety monitoring was performed by Data Monitoring Committee.
Data processing was completed by screening for out-of-range data, with cross-checks for conflicting data within and between data collection forms by a data manager. A random 10% of the data were independently validated against the source documents by the data manager and found to have no errors.
Safety monitoring was performed by using a list of expected serious adverse events: (1) intraventricular hemorrhage defined as hemorrhage causing ventricular dilatation with or without brain parenchymal involvement, (2) periventricular leukomalacia on cranial ultrasound scan imaging, (3) necrotizing enterocolitis requiring surgery, (4) patent ductus arteriosus (PDA) requiring treatment, (5) retinopathy of prematurity requiring laser treatment, (6) pneumothorax within 7 days after extubation, (7) evidence of traumatic nasal injury, (8) pulmonary hemorrhage, and (9) death.
A total of 544 infants were randomized between June 2011 and December 2014 from 8 NICUs, with 270 in each arm in the ITT dataset after excluding 4 infants who were randomized in error. The median (range) number of patients recruited at each site was 69 (4–145). A total of 304 were in the <28 weeks’ gestation strata and 236 were ≥28 weeks’ gestation strata. All infants were ventilated after delivery, either according to clinical need due to respiratory distress syndrome (defined here as the early onset of respiratory distress requiring supplemental oxygen and respiratory support within 4 hours of delivery) or after intubation immediately after delivery for administration of surfactant. The participant flow was as shown in Fig 1. The commonest reason for major protocol breaches was partial pressure of carbon dioxide more than 7 kPa at the time of randomization. The commonest reason for loss to follow-up at 28 days and 36 weeks was death. Baseline characteristics were well balanced between the 2 groups as shown in Table 1. The condition of infants before extubation was also comparable, as shown in Table 2. The median (range) age of infants at extubation was 1 day (0–14).
Extubation failed in 57 (21%) of 270 of the n-BIPAP group compared with 55 (20%) of 270 in the n-CPAP group, giving an odds ratio of 1.01 (95% CI 0.65–1.56), P = .97. Sensitivity analyses using the per-protocol dataset and gave very similar results with extubation failing in 50 (19%) of 259 of the n-BIPAP group compared with 52 (20%) of 254 of the n-CPAP group, giving an odds ratio of 0.91 (95% CI 0.58–1.44), P = .68.
Overall 91 (30%) of 304 extubations failed in the <28 weeks’ gestation strata and 21 (9%) of 236 in the ≥28 weeks’ gestation strata.
The effect of treatment allocation on the secondary outcomes is shown in Table 3. There were no significant differences in any of the measures.
The effects of treatment allocation on the serious adverse events monitored were as shown in Table 4. Seventy-seven infants in the n-BiPAP group and 87 infants in the n-CPAP group had serious adverse events.
There was no significant difference in the treatment effect (odds of extubation failure rate of n-BIPAP compared with n-CPAP) between the 2 gestation strata (P = .997, interaction test). In the <28 weeks’ gestation strata 47 (30%) of 157 of the n-BIPAP group compared with 44 (30%) of 147 in the n-CPAP group failed extubation, giving an odds ratio of 1.01 (95% CI 0.61–1.67), P = .97. In the ≥28 weeks’ gestation strata, 10 (9%) of 113 of the n-BIPAP group compared with 11 (9%) of 123 in the n-CPAP group failed extubation, giving an odds ratio of 1.01 (95% CI 0.41–2.5), P = .99.
In this large multicenter randomized trial comparing n-BiPAP and n-CPAP at equivalent MAP in infants born before 30 weeks’ gestation and <2 weeks old, we found no significant differences in extubation failure rates at 48 hours post-extubation between the 2 groups. There were no significant differences in the secondary outcomes including extubation failure at 7 days, oxygen requirement at 28 days, and at 36 weeks’ corrected gestation. There were also no significant differences in the duration of hospitalization (total and post-extubation) and any of the serious adverse event outcomes monitored. The 95% CI of the almost 0 effect of treatment translates to an absolute difference in extubation failure rates of –5% to 7%, less than the clinically relevant effect size specified before the trial began (±10%).
A limitation of the trial is the lack of blinding, leaving a potential for bias. However, strict criteria for extubation failure were defined and monitored. Per-protocol analyses excluding data when protocol deviations occurred showed no difference in results. As failure of extubation was based on blood gas criteria rather than response of the clinician to the blood gas parameters, it was possible to be precise at the time at which the infant failed extubation attempt. Second, there is no overall consensus on the optimal weaning strategy from CPAP. Our weaning protocol was based on consensus that it makes physiologic sense to wean the MAP initially until the infant is stable on 4 cm water pressure as this avoids abrupt withdrawal of positive pressure and risk of atelectasis. Last, because we aimed to compare equivalent MAPs in both arms, it can be argued that the baseline pressure during biphasic support was lower than the continuous distending pressure given in the corresponding CPAP arm. However, the baseline pressure was maintained at or above 4 cm water and is unlikely to have given rise to atelectasis or contributed any disadvantage to the n-BiPAP arm.
There have been few published trials to date directly comparing n-BiPAP with n-CPAP in preterm infants. Migliori et al4 performed an unblinded crossover study comparing 4 alternating phases of n-CPAP and n-BiPAP in 20 infants (gestational ages 24 to 31 weeks) within 6 hours of weaning from mechanical ventilation but delivered MAP was higher with n-BiPAP. Significant improvements in oxygen saturations and transcutaneous partial pressure of oxygen and carbon dioxide were noted during the n-BiPAP phases. There was also a significant reduction in spontaneous respiratory rate during the n-BiPAP phases. O’Brien et al5 reported a randomized trial comparing n-BiPAP with n-CPAP after extubation in 136 infants with <1250 g birth weight by using a similar design to the present trial. This also showed no apparent benefit between n-CPAP and biphasic pressure for support postextubation. Lista et al6 conducted a randomized controlled trial of 40 infants comparing the use of n-BiPAP with n-CPAP in premature infants after intubation-surfactant-extubation approach. They showed a significant reduction in duration of respiratory support, duration of oxygen dependency, and gestational age at discharge in the group receiving n-BiPAP and no rise in inflammatory cytokines as markers for associated lung injury, concluding that n-BiPAP was safe and well tolerated in this population. Kirpalani et al7 conducted a randomized trial involving 1009 infants <30 weeks’ gestation and 1000 g birth weight comparing different methods of noninvasive ventilation (defined as any technique combining n-CPAP with intermittent increases in applied pressure, including but not restricted to n-BiPAP) with n-CPAP and showed no difference in the rate of chronic lung disease or the need for intubation between the 2 groups. However, data related specifically to the efficacy of n-BiPAP in preventing chronic lung disease or extubation failure were not available.7 More recently, Salvo et al8 reported a randomized trial of synchronized nasal intermittent positive pressure ventilation versus n-BiPAP for primary respiratory support in 124 infants <1500 g birth weight and <32 weeks’ gestation that showed no difference in the requirement for intubation, death, or bronchopulmonary dysplasia between the 2 groups.
This trial for the first time clearly demonstrates that n-BiPAP confers no significant advantage on preventing primary extubation failure within the first 14 days of life when compared with n-CPAP at equivalent MAP (≤6 cm water) in infants born before 30 weeks' gestation. The results of this trial need to be applied cautiously in the broader neonatal population and beyond 14 days of age. At a later stage in neonatal care in which there may be established parenchymal lung disease/evolving chronic lung disease of prematurity, lung mechanics are different with poor compliance and nonhomogeneous lung inflation. These infants are likely to require higher MAP and also may be at risk for uneven lung inflation with focal atelectasis along with areas of overdistension if continuous high MAPs are applied. It is possible that n-BiPAP could allow the delivery of MAP (>6 cm water) with less risk of focal overdistension compared with n-CPAP at an equivalent MAP and further trials comparing n-BiPAP with n-CPAP in this population are warranted.
In the present trial, n-CPAP and n-BiPAP were delivered by using the Infant Flow Advance (CareFusion–BD Inc, Yorba Linda, CA.). This is a CPAP driver that has settings for single-level or biphasic support via a distal interface comprising either nasal prongs or a nasal mask, both of which are sized appropriately for the individual infant and give consistent pressure delivery. Variable flow is generated by diversion of the inspiratory gases by the infant’s expiratory flow at the distal interface (Fluidic Flip). n-CPAP is most commonly delivered by using a variable-flow device and we believe our findings are likely to be directly relevant to current neonatal practice by using a comparable device for delivery of n-CPAP.
This trial provides clear and conclusive evidence that there is no clinically significant difference in extubation failure rates at 48 hours post-extubation between n-BiPAP and n-CPAP at equivalent MAP when used in preterm infants born before 30 weeks’ gestation and <2 weeks old.
We acknowledge the following for their contribution: Trial coordinators: C. Jennings, S. Khan, A. Hendrickson; Investigators at participating centers: N. Soni, M. Yadav, R. Gupta, A. El-Azabi, N. Maddock, B. Ofoegbu, C. Zipitis; Trial Steering Committee (independent members): J.D. Grainger, L. Webster, S. Rickard, L. Livingstone; Data Monitoring Committee members: S. Sinha, S. Gupta, S. Cotterill; Research Parent Group at St Mary’s Hospital, Manchester; Sponsor: Central Manchester University Hospitals NHS Foundation Trust; Research Network: Clinical Research Network for Children, Manchester
- Accepted May 11, 2016.
- Address correspondence to Suresh Victor, PhD, Centre for Developing Brain, Division of Imaging Sciences and Biomedical Engineering, King's College London, St. Thomas' Hospital, SE1 7EH, UK. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: This article presents independent research funded by the National Institute for Health Research under its Research for Patient Benefit Programme (grant reference PB-PG-0909–19171). The views expressed are those of the author(s) and not necessarily those of the National Health Service, the National Institute for Health Research, or the Department of Health.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Victor S; Extubate Trial Group
- Lista G,
- Casoldi F,
- Fontana P, et al
- Salvo V,
- Lista G,
- Lupo E, et al
- Copyright © 2016 by the American Academy of Pediatrics