Article Text

Download PDFPDF

Non-invasive positive pressure ventilation in the preterm neonate: reducing endotrauma and the incidence of bronchopulmonary dysplasia
  1. A A Hutchison1,
  2. S Bignall2
  1. 1Division of Neonatology, Department of Pediatrics, University of South Florida, Tampa, USA
  2. 2The Winnicott Baby Unit, Department of Paediatrics, St Mary’s Hospital, London, UK
  1. Professor A A Hutchison, Division of Neonatology, Department of Pediatrics, University of South Florida, 2A Columbia Drive, Tampa, FL 33606, USA; ahutchiz{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Forty years after its initial description, the clinical definition of bronchopulmonary dysplasia (BPD) has changed but its aetiology and treatment are still debated.1 ,2 The incidence of BPD is greatest in very low birth weight (VLBW) infants of <28 weeks’ gestation.1 The severity of BPD has decreased with advances in care, including surfactant treatment, but primary prevention of BPD by avoiding premature birth remains elusive. Targets for decreasing the incidence of BPD include reducing oxygen exposure, avoiding lung infection/inflammation and avoiding ventilator-induced lung injury.

Oxygen exposure is linked to the occurrence of BPD.3 Preterm human neonates have oxidant stress from birth, and animal data demonstrate poor antioxidant defences.4 ,5 Thus avoidance of unnecessary oxygen exposure from birth is recommended.6 ,7 However, BPD can develop when continuous supplemental oxygen is not administered.1

Fetal/neonatal inflammation, infectious and non-infectious, has been causally linked to BPD,1 ,2 but treatment with postnatal steroids is problematic. A reduction in BPD with inhaled nitric oxide may be limited to neonates with birth weights >1000 g.810

Although some neonates can develop BPD despite minimal mechanical ventilation, a significant association exists between invasive artificial ventilation and BPD.1 Indeed, room air artificial ventilation of the fetal lamb results in lung injury similar to BPD.11 Measures to minimise ventilator-induced lung injury must begin from birth.6 ,1214 Reducing the duration of artificial ventilation by early extubation to nasal continuous positive airway pressure (NCPAP) can minimise lung injury in immature baboons.15 However, extubation failure occurs with NCPAP mainly owing to apnoea,16 despite caffeine treatment that is associated with decreased BPD.17 Non-invasive positive pressure ventilation (NIPPV) increases successful extubation by ∼30%.1820 Although no strong evidence confirms that NIPPV reduces the occurrence of BPD,21 a 33% decrease in BPD may accompany the introduction of NIPPV.22

This review presents key physiological concepts relevant to NIPPV and clinical information relevant to its efficacy and application.


Working with the neonate to minimise lung injury

Importance and control of end-expiratory lung volume

We believe that when applying any mode of respiratory support the clinician should be aware of the neonate’s focus on end-expiratory lung volume (EEV). EEV is the subglottic lung volume at end expiration. Before birth the attainment and preservation of EEV is critical for fetal lung growth.23 Postnatally, the neonate’s chest wall and lung compliances differ from those of the adult, resulting in a lower relaxation volume (Vr) (fig 1).24 ,25 The newborn must breathe differently, raising EEV above the Vr mainly by controlling the coordinated actions of the laryngeal muscles and diaphragm.24 ,26 ,27 Breathing with this EEV minimises the energy expended, enhances surfactant function, ensures minimal pulmonary vascular resistance and optimises ventilation/perfusion and gas exchange.14 ,24 ,25 ,28 To preserve EEV, volume loss in expiration can be limited by post-inspiratory inspiratory diaphragmatic activity, by laryngeal narrowing, by starting the next inspiration before EEV can decay to Vr, and by reversing the order of onsets of laryngeal opening and inspiratory diaphragmatic activity.24 ,26 ,29 During normal spontaneous tidal breathing, the dynamic chest wall and lung volume–pressure (V–P) curves show that minimal pressure is applied to the epithelium and capillaries.25 When the upper airway is bypassed by an endotracheal tube (ETT), the coordinated laryngeal and diaphragmatic control of respiratory pattern and EEV is disrupted.30 Invasive artificial ventilation from a low EEV and with large tidal volumes leads to surfactant damage, decreased compliance and trauma.14 ,28

Figure 1 Volume–time diagrams of adult and neonatal lung volumes are shown. The dashed line in each tidal volume separates inspiration from expiration. At rest (relaxation), the end-expiratory lung volume (EEV) of the adult (top left) is termed the functional residual capacity (FRC). It is a relaxation volume (Vr). In the newborn (top right) the Vr is low in comparison with that of the adult. The neonate actively maintains an EEV above Vr. The grunting breathing pattern (bottom left) maintains lung volume above the EEV for variable periods of expiration before the volume returns to the EEV. The incremental pattern of breathing (bottom right) maintains lung volume above the EEV throughout expiration and, using laryngeal and diaphragmatic activities, achieves a true increment in lung volume.

The concept of endotrauma

“Endotrauma” is injury to the airways and lungs from the disruption of homoeostasis that occurs during, and sometimes after, artificial ventilation through an ETT (box 1). Laryngeal bypass and loss of control of lung volumes expose the neonate to the physical insult of invasive artificial ventilation through an ETT, with damage in experimental animals being detected initially by reduced expired nitric oxide.31 The pre-existence or development of inflammation and exposure to the toxic effects of oxygen contribute to subsequent abnormal pulmonary development.1 ,2 Thus endotrauma may directly lead to lung injury and also to a vicious circle of an increasing need for invasive care. Decreasing endotrauma by using non-invasive respiratory support can result in normal lung growth in premature animals15 ,32 and may result in less BPD.22 ,33 ,34

Box 1 Endotrauma

Pathophysiological consequences
  • Upper airway bypass

    • Disruption of adaptive breathing patterns

    • Loss of control of end-expiratory lung volume, tidal volumes and airway pressures

    • Inadequate humidification/excessive “rain out” of water into lungs

    • Decreased inspiration of nitric oxide owing to lack of that produced in the upper airway

  • Presence of endotracheal tube (ETT)

    • Increased resistance especially with secretions in ETT lumen or malposition

    • Stimulation of irritant receptors

    • Increased infection/inflammation

    • Exposure to plasticisers

    • Local trauma to larynx, trachea, nose and palate

    • Laryngeal oedema and vocal cord dysfunction

Clinical consequences
  • Upper airway bypass

    • Fall in oxygen saturations/bradycardia related to loss of end-expiratory lung volume/atelectasis

    • Altered pulmonary and systemic (cerebral) haemodynamics (pulmonary hypertension and increased risk of intracranial haemorrhage)

    • Promotion of ventilator-associated lung injuries with increased bronchopulmonary dysplasia (BPD)

    • Atelectatotrauma to airway epithelium with decreased expired nitric oxide

    • Absent response to inhaled nitric oxide with prolonged ventilation

  • Presence of ETT

    • Need for suctioning and saline installation

    • Irritation-related apnoea/ bradycardia/other cardiac arrhythmias/ hypertension

    • Blockage or malposition of ETT with increased resistance/hypoxia

    • Potential need for emergency reintubation that may be complicated by laryngeal oedema

    • Spontaneous breathing against resistance that may result in negative pressure oedema

    • Increased tracheitis/lung infection/inflammation potentially with increased BPD

    • Increased risk of systemic infection

    • Need for dexamethasone treatment for extubation

    • Stridor requiring reintubation

    • Post-extubation atelectasis with short-term vocal cord dysfunction after extubation

    • Long-term vocal cord damage and dysfunction affecting speech development

    • Subglottic stenosis/need for tracheostomy

    • Nasal, dentition and palatal abnormalities affecting speech development

    • Potential to increase BPD

Careful respiratory support and minimising endotrauma from birth

In parts of Scandinavia, delivery room management employs stabilisation, rather than aggressive resuscitation, and is viewed as important in attaining their low BPD incidence.12 ,35 ,36 The neonate’s establishment of EEV is observed, with NCPAP and oxygen provided only if required.12 ,35 High tidal volumes, delivered via bag and mask, are avoided since in animals these reduce lung compliance, and this persists after surfactant administration.12 ,35 Hypocapnia is avoided as it causes glottic closure37 and cerebral vasoconstriction. Endotracheal intubation is minimised by intubating only briefly to administer surfactant.13 The asphyxiated neonate with a low EEV at birth may breathe out against a closed glottis, promoting recruitment.38 ,39 After endotracheal intubation this manoeuvre is ineffective,30 producing a rejection reflex.38

Respiratory support, optimising EEV, minimising endotrauma and NIPPV

The first focus when applying neonatal respiratory support should be upon the EEV. Although there is no simple way to measure this objectively, and thus interobserver assessments can differ, EEV can be is monitored by observing chest wall expansion, the presence/absence of dyspnoea and the diaphragm position on the chest x ray examination. Loss of EEV during artificial ventilation produces oxygen desaturation40 and is treated by recruiting EEV and weaning the fraction of inspired oxygen (Fio2). Deciding when to extubate requires that the infant’s respiratory drive and overall status, including sedation/analgesia, be balanced with the continuing risks of an ETT. Once extubated, apnoea/hypopnoea may limit the efficacy of NCPAP16 such that the lungs are not “opened up and kept open”,14 with atelectasis and desaturations increasing the Fio2 requirement. Given these limitations, early NCPAP use may not decrease BPD.3 ,41 For the VLBW infant at risk of BPD, the natural extension of NCPAP is NIPPV as it provides ventilatory support1820 22 and interacts with the neonate’s coordinated laryngeal and diaphragmatic muscle activities that control EEV.42


Practical aspects of using NIPPV in newborns


All NIPPV devices must deliver controlled humidification and have appropriate safety features. The NIPPV interfaces are similar to those for NCPAP.1820 Prongs or nasal mask sizes are adapted for nostril size and growth. Application of NCPAP/synchronised NIPPV (SNIPPV) interfaces can cause disfigurement/infection,43 ,44 thus prongs/mask sites are checked assiduously. Facemask NIPPV that was secured with tightly applied head straps is no longer used.45 Short prongs have less resistance than longer nasopharyngeal ones,46 thereby improving thoracoabdominal synchrony during NIPPV.42 Variable-flow NCPAP is the best short prong device technically,4650 but this may not uniformly result in clinical advantage.16 ,51 A modified variable-flow NCPAP device (EME-Viasys, Warwick, UK) delivers SNIPPV, employing an abdominal pneumatic capsule to detect diaphragmatic descent. In theory, this ensures glottic patency before flow is triggered. In published reports in neonates, NIPPV and SNIPPV pressures have been applied via short and long prongs using a conventional ventilator.1820 52 53 A high-frequency interrupter ventilator has been used in lambs.54

Box 2 Potential contraindications to synchronised non-invasive positive pressure ventilation

  • Congenital airway and lung anomalies

  • Untreated surfactant deficiency

  • Shock/hypovolaemia/sepsis

  • Abdominal disease/distension

  • Haemodynamically significant patent ductus arteriosus

  • Decreased respiratory drive (severe apnoea/sedation)

  • Nasal trauma

Studies evaluating NIPPV in the newborn

Only SNIPPV is known to be effective in weaning neonates from invasive artificial ventilation.1820 22 37 55 56 The use of NIPPV, as opposed to SNIPPV, does not ensure that the pressure is applied in synchrony with glottic opening. We speculate that gastric rupture is a risk with NIPPV55 whereas this is not reported with SNIPPV.57 Adult studies show that ineffective ventilation can occur with NIPPV in contrast to SNIPPV.37 ,56 In neonates, SNIPPV reduces thoracoabdominal asynchrony, respiratory rate and the apparent work of breathing and enhances tidal gas exchange with lower carbon dioxide levels.58

Box 3 Potential benefits of synchronised non-invasive positive pressure ventilation

  • Maintenance of end-expiratory lung volume (airway patency)

  • Less “atelectatotrauma/ barotrauma/ volutrauma/ endotrauma”

  • Less bronchopulmonary dysplasia/hypocapnia

  • Less infection

  • Early weaning/less reintubation

  • Improved nutrition

  • Improved parental interactions

  • Less tracheal/vocal cord/palatal injury

Three controlled trials have compared SNIPPV with NCPAP after extubation in 159 preterm infants.1821 The VLBW infants were aged 1–3 weeks and met criteria for weaning from artificial ventilation (Fio2 ∼0.35; peak inspired pressure ∼16–20 cm H2O; positive end-expiratory pressure ∼5 cm H2O; rate 10–25/min). Aminophylline was given before extubation in one study.20 The SNIPPV was applied by a ventilator triggered by an abdominal capsule to deliver peak inspired pressure ∼16–20 cm H2O, positive end-expiratory pressure 4–6 cm H2O; and rate 10–20/min,1820 with the SNIPPV settings regulated by blood gas and clinical assessments. Weaning from SNIPPV was achieved by reducing the assisted breath numbers and applied pressures or by switching to NCPAP and reassessing.1820 Set criteria identified extubation failure before 48–72 hours (table 1). The rate of weaning from invasive ventilation improved by 29%,19 32%18 and 33%20 with SNIPPV compared with NCPAP, with an overall SNIPPV success rate of ∼91%.

Table 1 Criteria for extubation failure in VLBW neonates1820

Box 4 Potential risks of synchronised non-invasive positive pressure ventilation

  • Air leak

  • Gastrointestinal distension/perforation

  • Infection

  • Lack of training and audit

  • Increased nursing care required

  • Cosmetic sequelae

None of the weaning trials was designed to study the impact of SNIPPV upon the incidence of BPD (oxygen requirement at 36 weeks with characteristic radiographic changes). In two trials no difference in BPD was reported: 55% NCPAP vs 44% SNIPPV19; 53% NCPAP vs 35% SNIPPV,20 with the latter also reporting a decrease in retinopathy of prematurity: 57% NCPAP vs 32% SNIPPV (p = 0.08).20 In two trials SNIPPV rescue was offered to NCPAP extubation failures21 and this prevented reintubation in 6/718 and 9/12,19 respectively. In a case-controlled study comparing outcomes in neonates with a mean gestational age of 26 weeks treated with SNIPPV (n = 30) or NCPAP after extubation (n = 30), significant reductions in the duration of supplemental oxygen and the incidence of BPD (73% NCPAP vs 40% SNIPPV) followed the introduction of SNIPPV.22 Caffeine treatment was given before extubation in this trial and may contribute to the efficacy of non-invasive ventilation since the incidence of BPD decreases by ∼10% in ⩽1250 g neonates given caffeine before 10 days of age.17 It is speculated that the primary way in which caffeine treatment decreases BPD is through less exposure to respiratory support, especially positive airway pressure.17 We speculate that caffeine will promote interactions between spontaneous breathing patterns and SNIPPV, with resultant optimisation of EEV and respiratory mechanics42 and a reduced need for positive airway pressure and oxygen exposure in preterm infants.

Two randomised controlled trials have compared NIPPV and NCPAP treatments for apnoea of prematurity in infants receiving either aminophylline or theophylline.52 ,53 ,57 A crossover trial of NIPPV and NCPAP in 20 infants, with mean gestational and postnatal ages of 26 weeks and 25 days, found no difference in the hourly number of apnoeic events.52 The other trial of 34 preterm infants, with mean gestational and postnatal ages of 27 weeks and 15 days, reported a significant reduction in hourly apnoeic episodes with NIPPV.53 Neither trial was designed to study the need for intubation and only one infant was intubated for apnoea. Future NIPPV studies in severe apnoea are warranted using the newer SNIPPV mode.57


We expect that in neonates with respiratory distress SNIPPV will be applied with caution to those most likely to benefit from it. Clinical decisions will balance the potential contraindications, risks and benefits, listed in boxes 2–4. Finally, although physiological considerations and adult data indicate that NIPPV should be applied in the synchronised mode, no studies have compared NIPPV and SNIPPV in neonates.


This review emphasises that SNIPPV is part of a total package of care that extends from fetal life onwards, involving appropriate application of proven treatments and minimising exposure to known risks. The provision of respiratory support should reflect the neonate’s respiratory physiology, emphasising the neonate’s focus on EEV and the associated adaptive breathing patterns and the need to minimise “endotrauma”. The neonate can apply adaptive breathing patterns when extubated to non-invasive respiratory support. Treatment with NCPAP does not always ensure that the lungs are “opened up and kept open” or that hypoventilation due to apnoea is offset. The natural extension of NCPAP treatment is SNIPPV, which increases successful extubation and may reduce the incidence of BPD and perhaps that of retinopathy of prematurity. A trial of SNIPPV powered to determine its impact on the incidence of BPD is starting. It will be important to determine how SNIPPV affects growth and development.



  • Competing interests: Neither author has any conflicts of interest and specifically neither is involved with any active clinical or contractual research funded by any company manufacturing a NIPPV/SNIPPV device. SB was involved in the development of the EME-Viasys SNIPPV device.