Article Text
Abstract
Context Sustained inflation (SI) has been advocated as an alternative to intermittent positive pressure ventilation (IPPV) during the resuscitation of neonates at birth, to facilitate the early development of an effective functional residual capacity, reduce atelectotrauma and improve oxygenation after the birth of preterm infants.
Objective The primary aim was to review the available literature on the use of SI compared with IPPV at birth in preterm infants for major neonatal outcomes, including bronchopulmonary dysplasia (BPD) and death.
Data source MEDLINE, EMBASE and the Cochrane Central Register of Controlled Trials, until 6 October 2014.
Study selection Randomised clinical trials comparing the effects of SI with IPPV at birth in preterm infants for neonatal outcomes.
Data extraction and synthesis Descriptive and quantitative information was extracted; data were pooled using a random effects model. Heterogeneity was assessed using the Q statistic and I2.
Results Pooled analysis showed significant reduction in the need for mechanical ventilation within 72 h after birth (relative risk (RR) 0.87 (0.77 to 0.97), absolute risk reduction (ARR) −0.10 (−0.17 to −0.03), number needed to treat 10) in preterm infants treated with an initial SI compared with IPPV. However, significantly more infants treated with SI received treatment for patent ductus arteriosus (RR 1.27 (1.05 to 1.54), ARR 0.10 (0.03 to 0.16), number needed to harm 10). There were no differences in BPD, death at the latest follow-up and the combined outcome of death or BPD among survivors between the groups.
Conclusions Compared with IPPV, preterm infants initially treated with SI at birth required less mechanical ventilation with no improvement in the rate of BPD and/or death. The use of SI should be restricted to randomised trials until future studies demonstrate the efficacy and safety of this lung aeration manoeuvre.
- Infant
- Newborn
- Delivery room
- Neonatal Resuscitation
- Positive Pressure Respiration
- Sustained inflation
Statistics from Altmetric.com
What is already known on this topic
Sustained inflation (SI) at the onset of neonatal resuscitation results in faster increase in functional residual capacity compared with intermittent positive pressure ventilation alone.
Cohort studies report significant reduction in need for intubation and lower incidence of bronchopulmonary dysplasia using SI.
What this study adds
Preterm infants initially treated with SI at birth required less mechanical ventilation in <72 h after birth.
SI did not improve the rate of bronchopulmonary dysplasia and/or death.
Significantly more infants treated with SI received treatment for patent ductus arteriosus.
Introduction
Establishing breathing and improving oxygenation after birth is vital for survival and the long term health of preterm infants. However, when infants fail to breathe after birth, an international consensus recommends intermittent positive pressure ventilation (IPPV) along with a baseline positive end expiratory pressure (PEEP).1 The goals of IPPV are to establish functional residual capacity (FRC), deliver adequate tidal volume to facilitate gas exchange and stimulate breathing while minimising lung injury.2 At birth, the lungs of very preterm infants are uniquely susceptible to injury because they are structurally immature, surfactant deficient, fluid filled and not supported by a stiff chest wall.3 Hence the lung of extremely preterm infants is easily damaged by mechanical ventilation. Animal studies have demonstrated that sustained inflation (SI): (i) improves lung compliance without adverse circulatory effects;4 (ii) achieves lung aeration more uniformly;4 (iii) has an increased inspiratory volume and greater FRC compared with IPPV alone;4 (iv) stabilises neonatal cerebral oxygen delivery, possibly protecting against cerebral hyperoxia;4 and (v) does not cause overdistension of the lung.4
In case series of asphyxiated term infants, an SI of 5 s showed a twofold increase in FRC compared with IPPV alone.5 ,6 These observations are supported by an experimental non-breathing rabbit model, which demonstrated that an SI of 20 s followed by IPPV coupled with PEEP resulted in a rapid increase in FRC.4 ,7 SI and PEEP potentially have additive effects in improving lung mechanics and reducing the need for respiratory support.4 ,7 Two cohort studies treated preterm infants with an SI of 15 or 20 s and reported a significant reduction in the need for intubation, lower incidence of bronchopulmonary dysplasia (BPD) and intraventricular haemorrhage, and decreased length of hospital stay compared with IPPV with routine intubation.8 ,9
These studies prompted the launch of randomised control trials using SI and IPPV compared with IPPV alone at the initiation of respiratory support at birth. The primary aim of this article was to review the available literature on the use of SI compared with IPPV at birth in preterm infants for major neonatal outcomes, including BPD and death.
Methods
Search strategy
We searched PubMed, EMBASE and Cochrane Central Register of Controlled Trials until 6 October 2014 using a predefined algorithm (see online supplementary appendix 1), and reviewed abstracts from the annual meetings of the Paediatric Academic Societies (2000–2013) and the European Society of Paediatric Research (2000–2013). Search concepts included ‘infant’, ‘newborn’ and ‘sustained inflation’ (see online supplementary appendix 1). A manual search of the bibliography of the included studies was also performed.
Study selection
Two reviewers (GMS and MK) independently reviewed citations for selection. Studies were included in the review if they met the following criteria: randomised controlled trial; comparing SI versus IPPV as the primary respiratory intervention during respiratory support in preterm population <33 weeks’ gestational age in the delivery room; and presented the outcomes of either death at latest follow-up or BPD (defined as the need for oxygen support or mechanical ventilation at 36 weeks’ corrected gestational age) during hospital stay. Secondary outcomes included the need for any mechanical ventilation during hospital stay; need for surfactant treatment; pneumothorax; need for postnatal corticosteroid treatment; intraventricular haemorrhage (grade III/IV or described as severe); periventricular leukomalacia; necrotising enterocolitis; patent ductus arteriosus (ie, needing medical treatment and/or surgical ligation); and retinopathy of prematurity (ie, any stage or severe retinopathy of prematurity). Full articles for potentially relevant trials were retrieved and independently assessed for their eligibility using a standardised data collection form. We also aimed to identify and, if available, include multiple publications describing the same trial. Authors of published articles were contacted for clarification and additional information relevant to the review. Discrepancies regarding inclusion were resolved through consensus among the review team. Studies with interventions after admission to the neonatal intensive care unit were excluded.
Data extraction
Data were recorded using a standardised data collection form to record study design and methodological characteristics, patient characteristics, interventions and the outcomes presented. Information regarding randomisation mode, allocation concealment, blinding, incomplete outcome data and selective outcome reporting were also documented. We extracted the number of randomised patients and number of analysed patients for each treatment group of each trial, and event rates for binary outcomes. Data extraction was independently performed by two investigators (GMS, MK) and discrepancies were resolved in consultation with another member of the review team (P-YC).
Assessment of risk of bias
We assessed the risk of bias of the included trials using elements of the Cochrane Collaboration tool.10 The domains used in the present systematic review pertained to randomisation and allocation concealment (selection bias), blinding (performance and detection bias), incomplete outcome data (attrition bias) selective outcome reporting and other biases. Each study was assessed as high, low or unclear risk of bias on each component. GSM and MK independently reviewed each included study for the risk of bias assessment and discrepancies were resolved in consultation with another member of the review team (P-YC).
Statistical analysis
The principal summary measure was relative risk (RR) for all dichotomous outcomes selected. We also calculated absolute risk reduction (ARR) and corresponding 95% CIs for the assessed outcomes. Heterogeneity was explored using a χ2 test, and the quantity of heterogeneity was measured using the I2 statistic.11 If substantial (I2>50%) heterogeneity was detected, the potential causes for its existence were explored and further sensitivity analysis undertaken. We summarised data using random effects DerSimonian–Laird meta-analysis models.12 Analyses were performed in RevMan V.5.2 (Cochrane Collaboration, 2013). The numbers needed to treat were calculated for all outcomes where the pooled estimates of RR were statistically significant. We planned a priori sensitivity analyses as per gestational age (for subjects <29 weeks’ gestation), duration of SI used in each trial and a subgroup analysis based on devices used for SI. The study is reported according to the PRISMA checklist.13
Results
Figure 1 shows the flow of studies through the selection process. Risk of bias assessments of the included trials is presented in table 1. The majority of the trials (three trials) scored unclear for sequence generation as the method of random sequence generation was not adequately explained and all trials used opaque envelopes for the purpose of allocation concealment. However, there was some risk of bias in the trial by Lindner et al14 due to the small fixed block size used for randomisation. Blinding of interventions from caregivers was not possible due to the nature of the interventions, but Lista et al15 stated that they tried to limit bias by blinding the outcome assessors to study interventions. Two studies15 ,16 were recorded as low on selective outcome reporting as these had study protocols available online making those risk of bias assessments possible.
Our initial search identified 371 citations of potentially eligible studies of which 330 were rejected based on screening of the study title and abstracts. Of the remaining 19 studies that were assessed in full text, four trials, enrolling 611 infants, fulfilled the inclusion criteria (table 2).14–18 In addition, we identified a letter to the editor,19 which described data from infants <28 weeks from the te Pas et al trial.16 All studies were described as randomised; three studies randomised participants before birth15–17 and one study after birth.14 Harling et al17 used a factorial design comparing: (i) SI of 5 s duration with conventional lung inflation of 2 s duration for the first inflation at birth and (ii) 100% oxygen compared with 50% oxygen for resuscitation at birth. Lindner et al14 used an adaptive interim analysis after 30 patients to estimate the final sample size of 110. However, due to slow enrolment, this study was stopped after 61 patients were randomised. Although three studies stratified using gestational age at birth, data were not available for outcomes stratified by gestational age.14–16 All of the studies provided inhospital outcome data for all randomised patients. Infants in all studies were analysed by intention to treat; however, none of the studies masked the intervention or reported long term outcomes. The SI and IPPV groups were well matched, with no significant differences in baseline birth weight or gestational age (table 2). Additionally, the trials did not differ in the rates of caesarean section (data not shown). One study used a facemask,14 ,17 one study used a single nasal prong15 ,16 and one study used either a face mask or shortened endotracheal tube12 for SI and IPPV. Lista et al started in 21–40% oxygen depending on the participating centre,15 two studies started in 100% oxygen14 ,16 and Harling et al17 ,18 randomised to either 50% or 100% oxygen. Other aspects of respiratory treatment, including resuscitation devices used, criteria for endotracheal intubation and surfactant administration, were adequately described in all studies and conformed to current practice.
Table 3 and figure 2A–C show pooled results from the trials. The pooled analysis showed no difference in BPD rates for babies treated initially with SI (RR 0.84 (0.57 to 1.22), ARR −0.05 (−0.15 to 0.05), I2=45%) (figure 2A). Similarly, there was no difference in the outcome of death (RR 1.39 (0.79 to 2.46), ARR 0.03 (−0.03 to 0.09), I2=0%) (figure 2B) or in the composite outcome of death or BPD among survivors (RR 0.92 (0.66 to 1.29), ARR −0.02 (−0.14 to 0.09), I2=50%) (figure 2C).
For the outcome of need for mechanical ventilation, the pooled analysis showed significant reduction in need for mechanical ventilation, as measured within 72 h after birth (RR 0.87 (0.77 to 0.97), ARR −0.10 (−0.17 to −0.03), number needed to treat 10) (figure 3). However, there was also a significant increase in medical or surgical treatment for patent ductus arteriosus in infants treated with SI (RR 1.27 (1.05 to 1.54), ARR 0.10 (0.03 to 0.16), number needed to harm 10) (figure 4) and a trend towards higher rates of intraventricular haemorrhages (RR 1.59 (0.83 to 3.03), ARR 0.03 (−0.01 to 0.06)) (table 3, figure 5). We did not find any difference in the rate of other neonatal outcomes (table 3).
In sensitivity analyses, when the data from Harling et al17 were excluded (as the two groups were somewhat similar in the delivered intervention), the results for the outcomes of death, BPD, combined outcome of BPD/death or mechanical ventilation <72 h did not change.
We were able to include data from three trials in a sensitivity analysis examining gestational age <29 weeks. Data for <29 weeks were assessed separately; the results were not significant for any of the outcomes studied (figure 6A–D, online supplement). We also planned to carry out assessments of reporting biases using funnel plots and to explore heterogeneity using subgroup analysis, but due to data limitations we were unable to perform these analyses. We had also planned subgroup analyses according to the devices used for SI. However, there were not enough studies using each type of device to conduct subgroup analysis and reach any meaningful conclusion.
Discussion
The results of this meta-analysis demonstrate that providing SI to preterm infants in the delivery room results in beneficial short term respiratory effects with significantly lower numbers of infants mechanically ventilated within the first 72 h of life. There was no significant difference noted in the outcomes of BPD, death or the combined outcome of death or BPD among survivors. There was an increase in the number of infants requiring either medical treatment or surgical ligation of a patent ductus arteriosus. In addition, we observed a trend towards an increased risk of severe intraventricular haemorrhage in infants initially treated with SI (table 3).
The results of this review should be interpreted with caution for the following reasons. Only two studies15 ,16 had study protocols available online and thus were likely to be low risk for selective reporting. However, the remaining two studies (Lindner et al14 and Harling et al17) also uniformly provided data for the majority of important neonatal outcomes of interest for this systematic review. The included studies used different oxygen concentrations for the purpose of resuscitation. The study by Lista et al15 started in 21–40% oxygen, while two studies were performed in the era of 100% oxygen,14 ,16 and Harling et al17 used a factorial design and randomised infants to either 50% or 100% oxygen. A recent randomised trial comparing 21% with 100% oxygen during initial resuscitation reported lower oxygen injury markers and significant reduction in BPD.20 The threshold for intubation and mechanical ventilation varied across the ,studies from 40% to 60% oxygen,14–16 potentially influencing the risk of ventilator associated lung injury. In addition, the included studies varied in terms of the immaturity of the subjects. Only one of the trials enrolled infants of 23 weeks’ gestational age17 while the remaining trials included infants of ≥25 weeks’ gestational age. These extremely premature neonates represent a very high risk group with a high mortality and a very high need for early intubation in the delivery room. However, a recent meta-analysis of non-invasive versus invasive respiratory support in the delivery room reported improved outcomes for infants not intubated in the deliver room.21 Indeed, those infants at the limit of viability do have developed pulmonary structure, which is arrested at the developmental stage. It would be doubtful if the prevention of further lung injury would alter the course in the development of BPD. Different ventilation devices and strategies (eg, different range of pressure, duration and single vs multiple SI) were used to deliver IPPV or SI in each trial.14 ,16 ,17 Harling et al17 used a shorter SI duration than that demonstrated to have benefits in animal models. The intervention studied by te Pas et al16 included several elements, of which SI was just one. Consequently, it is not possible to determine how much, if any, of the differences observed between the groups was due to the use of SI.16
Our meta-analysis showed that infants treated with an initial SI required significantly more medical or surgical treatments for patent ductus arteriosus (RR 1.27 (1.05 to 1.54), ARR 0.10 (0.03 to 0.16), with a number need to harm of 11) (figure 4). A recent animal study reported that SI and IPPV produced similar rises in pulmonary artery blood flow, but differences in the right ventricular–ductus arteriosus contributions to this flow. These findings suggest that different perinatal circulatory adjustments accompany SI and IPPV lung aeration strategies.22–24 Another possible explanation for this finding could be that allowing early FRC establishment and associated reduction of pulmonary vascular resistance might induce rapid development of left to right shunting through the ductus. It is uncertain if this early change in ductal haemodynamics would lead to an increase in treatments for patent ductus arteriosus, despite the variation in treatment criteria between studies. In addition, we observed a trend towards an increase in intraventricular haemorrhage in the SI group (table 3) but long term neurological development remains unclear as no study has yet reported this outcome. Animal studies demonstrated that tidal volumes of approximately 20 mL/kg are delivered with each SI.22 High tidal volume delivery during IPPV at birth has been reported to cause cerebral haemodynamic disturbances and an increase in brain injury.23 Animal models reported that cerebral blood flow only transiently changes during SI24 but venous return might be impaired, which could explain the trend to higher intraventricular haemorrhage. Taken together, the potential haemodynamic effects of SI in regional circulations and therefore the adverse complications in these developing humans warrant caution and further investigations of its role in the respiratory care of preterm infants.
Implication for research
The optimal inflation time and inflation pressure(s) required to establish an effective FRC in infants requiring respiratory support have not been determined. In addition, future studies should examine if delayed cord clamping or cord milking, compared with earlier cord clamping, affects outcomes in the presence or absence of SI. Future studies need to address how ascertainment is affected by consent and selection. Ideally, all infants should be enrolled by deferral of consent, thus avoiding the criticism that babies that might benefit most from SI are never enrolled.
Three randomised trials are currently underway: Schmölzer et al are investigating different inflations times to achieve lung aeration; Urlesberger et al are examining changes in cerebral blood volume; and Kirpalani et al's study will include long term outcomes.25–27 All will address some of the above confounders. It will also be interesting to examine if similar effects on mortality and morbidity are observed with the current practice of using 21–40% oxygen instead of 100% oxygen to decrease oxidative stress and related pulmonary injury.20 ,28
Implication for clinicians
The studies included in this meta-analysis were somewhat heterogeneous in terms of starting oxygen concentration employed, device used to deliver SI and the duration of SI. These differences should be taken into account when interpreting the results of this meta-analysis. SI improved short term respiratory outcomes (ie, the need for mechanical ventilation within 72 h of birth) but neither BPD nor death (or their combined outcomes) was significantly altered when SI was delivered prior to IPPV to preterm infants. No study has yet reported long term outcomes. We therefore suggest that the use of SI in preterm infants should be limited to clinical trials until further studies demonstrate the efficiency and safety of this lung aeration manoeuvre in both the short term and long term.
Conclusions
While the meta-analysis supports the notion that SI improves FRC and therefore the need for mechanical ventilation during the first 72 h, the pulmonary protective effect is lost in the subsequent development to BPD, which has a multifactorial pathogenesis in addition to the heterogeneity of the studies. Further studies are needed to evaluate the efficacy, including long term outcomes and safety, of this lung aeration manoeuvre at birth.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online Appendix
- Data supplement 2 - Online figures
Footnotes
http://journals.bmj.com/site/authors/editorial-policies.xhtml#copyright
Contributors Collection and assembly of the data: GMS, MK, and P-YC. Analysis and interpretation of the data: GMS, MK, and P-YC. Conception and design, drafting of the article, critical revision of the article for important intellectual content and final approval of the article: GMS, MK, MOR, KA, GP, GL, and PYC. The corresponding author had full access to all of the data in the study and had final responsibility for the decision to submit for publication.
Funding GMS is a recipient of the Heart and Stroke Foundation/University of Alberta Professorship of Neonatal Resuscitation, and a Heart and Stroke Foundation Canada Research Scholarship. MOR is supported by a Molly Towell Perinatal Research Foundation Fellowship. The sponsor of the study had no role in study design, data collection, data analysis, data interpretation or writing of the report.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Fantoms